An evaluation of a temperament-focused parent-training program

MISSING IMAGE

Material Information

Title:
An evaluation of a temperament-focused parent-training program
Physical Description:
v, 124 leaves : ill. ; 29 cm.
Language:
English
Creator:
Sheeber, Lisa B., 1962-
Publication Date:

Subjects

Subjects / Keywords:
Parents -- education   ( mesh )
Temperament   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1991.
Bibliography:
Includes bibliographical references (leaves 117-123).
Statement of Responsibility:
by Lisa B. Sheeber.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 26253124
ocm26253124
System ID:
AA00011167:00001

Full Text












AN EVALUATION OF A TEMPERAMENT-FOCUSED
PARENT-TRAINING PROGRAM

















BY

LISA B. SHEEBER


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY



UNIVERSITY OF FLORIDA


1991














ACKNOWLEDGMENTS

Among the many who have contributed to this project, I

would like to thank first and foremost my chairperson, Dr.

James Johnson, whose knowledge, guidance, and patience has

been indispensable and much appreciated. I would also like

to thank my committee members, Drs. Sheila Eyberg, Hugh

Davis, Robin West, and Constance Shehan, for their direction

and support throughout this project. I similarly, would

like to express my appreciation to the undergraduate

research assistants who provided valued assistance in the

completion of this project and to my husband, Erik Sorensen,

for his statistical consultation and support. Additionally,

I would like to thank Tualatin Valley Mental Health Center

and Morrison Center for the generous provision of space in

their buildings. Finally, I would like to thank both the

preschool directors who provided access to parents of

preschoolers and the parents themselves whose active and

good-humored participation was truly essential.

















TABLE OF CONTENTS


ACKNOWLEDGEMENTS. . .

ABSTRACT. .

INTRODUCTION .

Child Temperament . .
Child Adjustment and Parent-Child Interactions
Maternal Adjustment and Family Functioning. .
Temperament-Based Parent Guidance. .
Specific Aims .


METHODS .


Subjects .
Measures .
Procedure.


RESULTS .

Homogeneity of Sample. .. .
Equivalence of Subjects across Conditions
Program Effectiveness. .
Clinical Impact of Group Participation .

DISCUSSION .

APPENDIX A UNPUBLISHED MEASURES .

APPENDIX B RAW DATA. .

APPENDIX C OUTLINE OF PARENT-TRAINING PROGRAM

REFERENCES .

BIOGRAPHICAL SKETCH . .


25

S 25
S 29
S 41

47

S 47
S 48
S 50
S 60

. 63

S 76

S 83

S 93

117

124


iii


Page

ii

iv

1

1
4
7
11
23


. .


.....



















Abstract of Dissertation Presented to the Graduate School of
the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

AN EVALUATION OF A TEMPERAMENT-FOCUSED
PARENT-TRAINING PROGRAM

By

Lisa B. Sheeber

August 1991

Chairperson: James H. Johnson, Ph.D.
Major Department: Clinical and Health Psychology

Child temperament has received increasing attention as

a factor relevant to child development, parental adjustment,

and family functioning. Due to the observed relationships

between difficult child temperament and poorer functioning

in each of these domains, there has been a call in the

literature for temperament-based parent guidance. The goal

of such programs is to improve the fit between parenting

styles and children's temperament characteristics in order

to alleviate the aforementioned difficulties. To date,

however, empirical investigations of these programs are both

sparse and methodologically weak. The current study

examined the efficacy of a temperament-based,

psychoeducational program for mothers with temperamentally

difficult preschoolers.

The results indicated that relative to subjects in a

wait-list control group, mothers who participated in the










parent-training group demonstrated improvements in affect,

satisfaction with the parent-child relationship, and

perceived parenting competence. Additionally, reductions

in mother-rated child behavior problems were also observed.

However, spousal relationships, feelings of being restricted

by the demands of parenting, and father-rated child behavior

did not demonstrate similar improvement. These results

provide preliminary support for temperament-based parent-

training.














INTRODUCTION

Child Temperament

Attention to child temperament as a factor relevant to

the understanding of child development and family

interactions has grown among both researchers and clinicians

(Bates, 1987; Carey, 1990; Rutter, 1987). Recent interest

in child temperament gained its impetus from the clinical

research of the New York Longitudinal Study (Thomas, Chess,

& Birch, 1968). These investigators defined temperament as

the stylistic component of behavior. They identified nine

dimensions of behavior which comprise child temperament:

(a) activity level, (b) rhythmicity/regularity in biological

functions, (c) approach vs withdrawal in response to new

stimuli, (d) adaptability to new or changed situations,

(e) threshold of responsiveness to stimuli, (f) intensity of

reaction, (g) quality of mood,(h) distractibility, and (i)

attention span/persistence. From five of these dimensions

(regularity, approach/withdrawal, adaptability, intensity,

and mood), three constellations of temperament were

discerned: difficult, easy, and slow to warm up. These

temperament characteristics and patterns are significant in

that they have implications for both the ease with which

children will adjust to their environment and the demands

that caring for them will place on their parents. It is










notable that a great deal of research has been spawned by

the Thomas and Chess conceptualization of temperament

(Black, Gasparrini, & Nelson, 1981; Chess & Korn, 1970;

Graham, Rutter, & George, 1973; Maurer, Cadoret, & Cain,

1985; McNeil & Persson-Blennow, 1982; Maziade, Caron, Cote,

Boutin, & Thivierge, 1990).

Congruent with the expansion of interest in child

temperament, however, the field has grown such that the

literature currently features several prominent working

models of child temperament (Bates, 1987; Goldsmith et al.,

1987). This diversity of models is perhaps not surprising

given the breadth of areas spanned by temperament

researchers. As summarized by Bates (1987), these range

from primarily biological foci as is seen for example, in

Goldsmith and Campos' (Goldsmith et al., 1987) work on

individual differences in emotional development through to

the study of social processes such as parent-child

interactions and the development of psychopathology (e.g.,

Chess and Thomas, 1984).

While maintaining sufficient similarity to fall under

the conceptual rubric of temperament theories, the various

models differ in several regards. Primary among these is

the emphasis placed on biological and in particular,

genetic, underpinnings of behavior. For example, while Buss

and Plomin (1984) consider heritability to be critical to

their conceptualization this is not true of all researchers

(e.g., Campos and Goldsmith, cited in Goldsmith et al.,









1987). Similarly, while documenting the biological

substrates of behavior is central to some theories (e.g.,

Rothbart .and Derryberry, 1981) others have sufficed to make

reference to the assumed constitutional basis of temperament

characteristics (Thomas & Chess, 1977). The salience placed

on the stability of temperament characteristics is another

central area of disagreement; while some consider stability

to be essential to the definition (e.g., Buss & Plomin,

1984) others have concluded that to the extent that behavior

is multi-determined, it is only reasonable to expect that

the expression of temperament characteristics will be

altered as a function of environmental demands (Chess &

Thomas, 1984). Finally, only limited consensus exists

regarding the specific behavioral dimensions which define

temperament; while most conceptualizations include the

characteristics of activity level and emotionality, other

dimensions vary across theories.

However, while even this very brief description

illustrates the diversity of theoretical conceptualizations,

most of the disagreement reflects a divergence on emphasis

rather than content. In a general way there is consensus

around several defining features of child temperament. As

summarized by Goldsmith et al. (1987) there is agreement

that temperamental dimensions reflect behavioral tendencies,

address individual differences in behavior, are likely to

have at least some biological component, and are expected to

demonstrate a greater level of continuity than other aspects










of behavior. Additionally, while most researchers

conceptualize temperament as theoretically being a "within-

the-person" variable, there is recognition that as suggested

by Rothbart (1982), temperament may more accurately be

conceptualized as reflecting a "within-the-[person]-within-

the-home" characteristic. This interpretation takes into

account both the pragmatic consideration that assessment of

temperament will be influenced by care-taker handling

(Rothbart, 1982) and the more far-reaching possibility that

the expression of temperament itself, may be modified by

environmental factors (Plomin, 1984).

Child Adjustment and Parent-Child Interactions

A relationship between difficult temperament

characteristics and child behavior problems, both

concurrently and predictively, has been well-documented

(Biederman et al., 1990; Cameron, 1977, 1978; Chess &

Thomas, 1984; Earls & Jung, 1987; Johnson, Basham, & Gordon,

1982; Webster-Stratton & Eyberg, 1982; Wolkind & De Salis,

1982). A broad array of difficulties has been noted in

children with difficult temperament characteristics

including depression, anxiety, sleep disturbances,

hyperactivity, and oppositional behavior styles. The

literature suggests that several temperament characteristics

may be operative in predicting child psychopathology. In

particular, low distractibility/high persistence, low

adaptability, low rhythmicity, and high intensity have been

identified as substantial contributors (Earls, 1981;










Garrison, Earls, & Kindlon, 1984; Thomas & Chess, 1977).

Other researchers have suggested that negative mood and high

activity.level/high distractibility may also be contributory

(Terestman, 1980; Thomas & Chess, 1977; Turecki & Tonner,

1985; Webster-Stratton & Eyberg, 1982). Clearly, the

specific temperament characteristics identified as

predictive of behavior problems has varied across studies.

It is likely that this a function of both the nature of the

assessment measures and the typology of behavior problems

being studied. In fact, recent studies have begun to

document specificity in the relationships between

temperament characteristics and the nature of behavior

disorders (Bates, 1990; Maziade et al., 1990).

It has been suggested that the relationship between

difficult temperament and behavior problems may be mediated

by the quality of parent-child interactions (e.g., Chess &

Thomas, 1984). This is notable in that significant to the

concept of child temperament is the implication that

children'possess traits that will influence both the way

they respond to the world and the way others, particularly

their parents, will respond to them. Numerous studies have

demonstrated that both the quantity and quality of parental

interaction with children are related to child temperament

characteristics (Campbell, 1979; Chess & Thomas, 1984;

Frankel & Bates, 1990; Gordon, 1983; Hinde, Easton, Meller,

& Tamplin, 1982; Klein, 1984; Stevenson-Hinde & Simpson,

1982). For example, one study (Webster-Stratton & Eyberg,










1982) found that mothers of active, low attention children

demonstrated more negative affect as well as more non-

accepting and submissive behavior than did other mothers.

Moreover, Lee and Bates (1985) demonstrated that toddlers

with difficult temperament characteristics were more likely

to have conflictual interactions with their mothers. These

investigators proposed that these interactions, similar to

the coercive pattern described by Patterson (1982), may be a

manifestation of the proposition that conflict between

parent and child due to mismatch between child temperament

and parental rearing techniques could increase the

likelihood of later behavior problems.

Several theorists have suggested that it is this

"goodness of fit", or the "match" between the child's

characteristics and those of the parents, which will

determine the level of stress in the interaction as well as

the potential for behavior problems (Belsky, Lerner, &

Spanier, 1984; Lerner, Baker, & Lerner, 1985; Thomas &

Chess, 1977). A good match is construed as one in which the

demands of the environment are congruent with the child's

capabilities and sufficiently flexible to accommodate his or

her difficulties. Variance in parental response to the

child will be reflected in his or her development and may

serve to either alleviate or aggravate the potential for

behavior problems (Cameron, 1978; Chess & Thomas, 1984;

Maziade et al., 1985; Thomas & Chess, 1977). For example,

Korn and Gannon (1983) have reported that difficult child










temperament was less predictive of behavior problems in a

working-class Puerto Rican sample than in an age and sex

matched sample of white, middle class families; these

investigators suggested that these differences reflected the

differential demands made on the children in the two

cultures.

However, it should be noted that a good match does

not imply a one to one correspondence between parent and

child temperament. For example, the interaction between a

parent and child both of whom have intense, negative moods,

would not have the promise of being particularly smooth.

Given a parent who's behavior is rigid, demanding or

inconsistent, an easy child may be more able than a

difficult child, to adapt without major strains. A

difficult child who is less regular and slower to adapt may

require more parental flexibility. An additional point is

that a good fit should not be construed to entail the

absence of stress. It is the excessive stress produced by a

poor fit'which is considered to be problematic within this

model (Thomas and Chess, 1977).

Maternal Adjustment and Family Functioning

Although the relationship between parent and child has

been described as one of reciprocal influence, the work

cited thus far has focused primarily on the child's

development. In line with the interactionist perspective

which emphasizes the bidirectional nature of influence,

several investigators have begun to focus on the










relationship between child temperament and parental

adjustment. In particular, it has been suggested that

parents of difficult children may experience a variety of

interpersonal and psychological difficulties.

The original work in this field was conducted by Thomas

and Chess (1977). Based on clinical contact with parents

and children as part of the NYLS, these investigators

elucidated a range of difficulties that parent of

temperamentally difficult children appeared to experience.

These included feeling threatened, anxious, and inadequate

as parents as well as being intimidated by the child's

behavior, resentful of the demands of parenting, and blaming

of the child. These researchers suggested, moreover, that

the relationship between child temperament and parental

adjustment may be mediated by the parents own

characteristics: parents whose own temperaments render them

less adaptable may be more vulnerable to these experiences.

Several studies have begun to delineate the

psychological difficulties that differentiate parents of

difficult children from their cohorts with easier children.

The most consistent findings in the earlier literature in

this area, appear to be an increased incidence of depression

(Cutrona & Troutman, 1986; Ventura, 1982; Wolkind and De

Salis, 1982) and anxiety (Ventura, 1982; Wolkind and De

Salis, 1982) in mothers of difficult infants. Decreased

marital satisfaction was also noted by Wolkind and De Salis

(1982). Studying a sample of mothers with preschool









children, Stevenson-Hinde and Simpson (1982) found that

child temperament was related to both anxiety and

irritability but not to depression.

Recently, Sheeber and Johnson (1991a) provided a more

in depth analysis of the relationship between child

temperament and maternal adjustment. These investigators

reported that mothers of temperamentally difficult preschool

children demonstrated more psychosocial difficulties than

did mothers of easier children. In accordance with the

majority of the previously cited work, the results indicated

that these mothers were more prone to anxiety and

depression. Additionally, they reported experiencing more

doubts regarding their competence as parents, feeling less

emotional closeness to their children, and finding parenting

to be more restrictive. Difficult child temperament was

also found to be related to more life style changes and

alterations of daily family routines. Relatedly, these

mothers reported greater marital discord especially as

regards the absence of emotional and active support in child

management. Contrary to expectation, the "mismatch" of

mother and child temperament was not found to be

additionally predictive of either maternal or familial

difficulties.

The quality and intensity of child mood emerged in this

study as the temperament characteristics most predictive of

maternal and familial difficulties. High activity level,

low distractibility, and irregularity were also clearly










operative. Excepting activity level, these dimensions are

consistent with those found in the earlier studies

(Stevenson-Hinde & Simpson, 1982; Ventura, 1982; Wolkind &

De Salis, 1982). Turecki (Turecki & Tonner, 1985) reported

that the combination of high activity level and high

distractibility was also related to maternal difficulty.

The reliance on maternal reports for obtaining indices

of both child temperament and maternal/familial adjustment

is a potential limitation of the work in this area. In

particular, the issue arises of the degree to which maternal

adjustment and temperament may bias child temperament

ratings so as to account for the significant findings.

However, this potential bias was controlled for

methodologically in several of these studies via either pre-

study assessment of maternal adjustment (Cutrona & Troutman,

1986; Wolkind & De Salis, 1982) and/or the utilization of

alternative measures of temperament (Cutrona & Troutman,

1986; Sheeber & Johnson, 1991a). The results therefore

suggest that a relationship exists between child temperament

and maternal adjustment independent of the potential

contribution of distortions in temperament ratings due to

maternal maladjustment.

Finally, it has been suggested that these parental

difficulties will be fed back into the parent-child system,

interfering with successful parenting and serving to

perpetuate a cycle of dysfunctional interactions. Thomas

and Chess (1977) maintain that a parent who is experiencing










any of the previously described difficulties cannot provide

the patient, gradual exposure to new situations and demands

that a difficult child requires. They are more likely to

pressure, appease or vacillate, all the while expressing

feelings of hostility, impatience, and confusion. It is

felt that this can lead to increasing difficulty on the part

of the child. These cautions become more potent when

considered in conjunction with the previously discussed data

regarding parent-child interactions, findings that child

temperament interacted with maternal depression (Wolkind &

De Salis, 1982) and maternal role-satisfaction (Lerner &

Galambos, 1985) in contributing to the development of

behavior problems, and evidence suggesting that difficult

children are more likely to be victims of parental abuse

than are their easier cohorts (Johnson, Floyd, & Isleib,

1983).

Temperament-Based Parent Guidance

In light of the observed relationships between child

temperament and a broad range of difficulties in the areas

of parental adjustment, parent-child interactions, and the

development of child behavior problems several investigators

have recommended the development of temperament-based parent

guidance programs for parents of temperamentally difficult

children (Cameron & Rice, 1986; Carey, 1982; Thomas & Chess,

1977; Turecki & Tonner, 1985). These investigators suggest

that such programs may serve to alleviate and/or prevent the

development of the aforementioned problems. In particular,










it has been proposed that such programs could reduce the

level of negative parental affect and facilitate more

effective and positive parent-child interactions.

Additionally, guidance could serve to reduce problem child

behaviors that may have developed, in part, as a function of

a poor fit between parenting style and child temperament

characteristics.

The conceptual framework for temperament-focused parent

training programs is rooted in the goodness-of-fit model

described earlier. In accordance with this model, the

primary goal of such programs is to reduce the excessive

stress produced by a poor fit between environmental demands

and child capabilities. As described by Lerner, Baker, and

Lerner (1985) poorness of fit can be addressed by

interventions directed either at the individual or

contextual level. These investigators concur with Chess and

Thomas (1984) in suggesting that in working with young

children, the context, in the form of parental expectations,

behavior, and attributions are the most viable target of

intervention. To the extent that it is feasible, however,

many clinicians advocate helping the child to recognize

their behavioral patterns and develop techniques for

adapting to situational demands (Chess & Thomas, 1986;

Turecki & Tonner, 1985).

As originally formulated by Thomas and Chess (1977),

temperament-focused parent training has two components.

Initially, the focus is on acquainting parents with the










notion of temperament under the assumption that an

appreciation of temperament as reflecting inborn

characteristics of the child would serve to reduce their

negative feelings and allow them to respond to their child

with less anger and apprehension. During the second phase,

the implications of temperament for child management are

explored. In particular, training focuses on developing

behavior management strategies which are congruent with the

child's temperament characteristics. This general framework

has been adopted to varying degrees by other clinicians

although the extent to which both components are

incorporated is variable.

At the present time, only a handful of temperament-

based parent guidance programs have been reported in the

literature. The first of these was described in the early

work of Thomas, Chess, and Birch (1968). In this project,

parent guidance was utilized for 42 cases in which a

mismatch between child temperament and parenting styles was

thought to be responsible for the development or maintenance

of behavior problems. It is notable that this sample was

not limited to difficult temperament children but rather

included children demonstrating a range of temperament

patterns. While the report of the program is largely

anecdotal, several key features were emphasized. Primary

among their recommendations is the necessity of an

individualized program in which the details of parent and

child behaviors contributing to the poor fit are identified










and alternative behaviors suggested to the parents.

Secondly, parents are encouraged to distinguish temperament

characteristics from parental attributions; for example,

lack of persistence is not synonymous with laziness.

Relatedly, parents are reassured that their parenting

approach is not "bad" but just incongruent with this child's

particular style. Finally, parental feelings associated

with parenting difficulty are discussed.

The investigators reported an overall success rate of

50% for this sample and indicated that the degree of

improvement was related both to the amount of behavioral

change exhibited by the parents and to the child's

temperament pattern. Slow-to-warm-up, easy, difficult, and

persistent children demonstrated 100, 70, 60, and 33%

improvement respectively. Distractible-nonpersistent

children did not improve. The investigators attributed this

latter finding to the reticence of urban middle-class

parents to accept this behavior pattern as it appears

incompatible with educational and occupational achievement.

Although the investigators are to be credited for their

innovative work in applying the findings of temperament

research to the alleviation of parent-child difficulties,

methodological limitations preclude clear interpretation of

their results. One major weakness in the design is the

failure to utilize statistical analyses for their between

group comparisons. It is, therefore, unclear whether the

differential effectiveness of parent guidance for children










with dissimilar temperament patterns reflects actual group

differences versus chance findings. This is a particular

danger given the small sample size for each temperament

pattern. For example, given that there were 10 difficult

and 10 easy children the difference between the 60 and 70%

effectiveness of treatment for the two groups is

attributable to one child. A related problem is that the

small sample size raises questions about the

generalizability of the findings: are the behavior problems

evidenced by distractible children actually less tractable

or were these four children and their parents less open to

change for reasons unrelated to the children's temperament

characteristics? Finally, as the original sample was

diverse as regards the types of behavior problems exhibited

it would have been interesting if the investigators had

examined whether some behaviors were more amenable to change

than others.

An additional difficulty with this study is their sole

reliance on clinical judgment for assessment of both initial

behavior problems and changes in parental and child

behavior. This is problematic both because of the potential

for examiner bias and because of the difficulty it presents

for the reader in attempting to understand the changes that

have occurred. Additionally, the absence of a control group

renders it difficult to interpret the behavioral changes as

necessarily a function of treatment. Although the

investigators report that their continuing clinical use of










parent guidance provides additional support for its efficacy

(Chess & Thomas, 1986) it is unfortunate that they have not

attempted to replicate their results in the two decades that

have passed since this original study was completed.

More recently, Turecki (Turecki & Tonner, 1985) has

expanded on this work from a solely clinical perspective.

He has detailed a treatment program specifically designed

for parents of temperamentally difficult children. Notable

in his conceptualization is the movement away from a

predetermined concept of difficultnesss" toward a definition

based on the parents' experience with the child. This

approach is more in line with the goodness-of-fit model as

it recognizes that different traits may be considered

difficult by parents with varying characteristics,

expectations, and values as well as in different situations

or developmental stages. For instance, distractibility may

be an asset in a young child whose attention can easily be

pulled away from dangerous objects but may become stressful

when the'older child does not complete homework or chores.

As described in The Difficult Child (Turecki & Tonner,

1985), the initial focus of the program is on providing a

definition of temperament, detailing its explanatory

function in understanding child behavior, and describing the

parental feelings and family problems that may be associated

with parenting a difficult child. The next part of the

program is designed to help the parents identify both their

child's specific temperament profile and the behavior










problems which may be expressions of it. The main goal of

this component is to facilitate an unemotional evaluation of

the child's behavior that is unencumbered by parental fears

or attributions. Subsequent to this, the parents are taught

a general approach to behavior management, techniques for

dealing with temperamental issues, and ways of dealing with

the family problems which may have, in part, developed as a

function of the child's behavior. Finally, this program is

supplemented by a parent support group to help the parents

deal with the negative feelings associated with problems in

parenting.

Presently, the remaining work in this area has become

increasingly focused on anticipatory guidance. This work

has been conducted in pediatric settings and has largely

involved parents of infants. Carey (1982) discussed three

ways in which the clinician can utilize temperament

information to foster improved parent-child interactions.

The least intrusive of these is an educational discussion in

which the clinician provides the parent with information

about normal individual differences. This approach,

congruent with the first phase in Thomas and Chess's (1977)

conceptualization, helps parents understand both that

disagreeable behavior is not necessarily indicative of

either poor parenting or child psychopathology and that the

application of standard approaches to child rearing will not

be equally successful with different children.










The second approach, recommended as a standard clinical

procedure, is to identify the child's temperament profile

via questionnaire techniques, interviews, and behavioral

observations. This information may provide the parent with

a more organized view of the child's behavior and thereby

facilitate interactions which are more congruent with the

child's temperament characteristics. Central to this

approach is the avoidance of labels in favor of descriptive

information when providing parents with feedback about their

child's temperament style. Utilizing a similar method,

although one based solely on screening with a temperament

questionnaire, Little (1983) reported that 75% of the

mothers receiving this information returned a brief survey

in which they were asked to rate the utility of the

information provided. Of these, 89.9% reported gaining a

better understanding of their child, 87.3% considered this

information to be worthwhile, and 57% reported that it

changed their parenting.

The final level of intervention is the provision of

guidance regarding alternative approaches to child

management when it appears that a poor fit is responsible

for stress in the parent-child interaction and the

development of reactive symptoms. Carey emphasizes that

intervention at this level requires the identification of

stress in the relationship and not merely a difficult

temperament profile; this is essential given that parental

expectations for and approach to coping with child behavior










will moderate the impact of the child's temperament

characteristics.

An alternative approach to utilizing temperament

questionnaire information obtained in early infancy has been

presented by Cameron and Rice (1986). Based on a review of

the literature, they identified what they considered to be

the most likely problems to arise at various stages of

infancy for children with different temperament profiles.

Written anticipatory guidance including a description of the

two problems most likely to occur, reasons for their

development, validation of the parental feelings likely to

be associated with these difficulties, and management

suggestions, was then mailed to the mothers. A notable

feature of this work is the investigators' attempt to

document the relationship between child temperament and

specific behavior problems. The results of two studies

indicated that various temperament profiles did predict the

occurrence of distinct behavior problems as reported by

mothers.' Unfortunately, the investigators did not report

the amount of variance predicted by this model.

Additionally, as only 48% and 67% of the mothers in the two

samples provided information regarding the occurrence of

problems it is unclear the extent to which this relationship

is sample specific.

With regard to the utility of the program, the results

of these studies were congruent with those reported by

Little (1983). Of the 48% of mothers in the first sample










who returned a follow-up questionnaire, 90% reported that

filling out the questionnaires was useful, 80% said the

guidance.material helped them to understand the behavior

problems that occurred, and over 70% found the handling tips

helpful. Perceived utility of the guidance materials was

related to both the number of temperament related problems

which occurred and the accuracy with which the emergence of

behavior problems had been predicted by the early

temperament data. The extent of these relationships is not

reported for the first issue and is relatively small (R=.13-

.27) for the latter.

Finally, the investigators report that guidance was

most effective for those difficulties related to energy

level in the fifth to eighth month (accident risk,

assertiveness, mealtime, and sleep issues) and suggest that

more extended counseling or peer support groups may be

necessary for parents whose infants continue to develop

temperament-related sensitivity and dependency issues in

late infancy. However, the criteria used for making this

evaluation are unclear.

Despite the apparent clinical utility of these

programs, several difficulties exist in this literature.

Firstly, methodological problems constrain the

interpretability of the obtained results. One major

limitation is the failure to adequately assess outcome. As

reported earlier, Carey's discussion of the utility of

temperament data did not address the need to demonstrate










effectiveness. While the other researchers (Cameron and

Rice, 1986; Little, 1983) have explored this issue, they

relied solely on maternal responses to broad surveys that

have unassessed psychometric characteristics. Additionally,

these investigators have not provided specific information

regarding the manner in which the guidance was beneficial in

either preventing the development or exacerbation of child

behavior problems or in alleviating maternal concerns.

Additionally, as the post-intervention assessments were

completed by mail, the extent to which the reported success

represents an overestimation based on a greater response

rate from individuals who felt the guidance to be beneficial

is unclear. Again, the absence of a control group in both

of these studies hinders attempts to interpret the findings.

In addition to the methodological difficulties,

concerns have been raised regarding the appropriateness of

general screening and anticipatory guidance for temperament

related problems. These concerns center around the risks

involved'in labeling a child "difficult" in light of the

limited utility of temperament ratings in early infancy for

predicting either later temperament characteristics or the

development of behavior problems (Rothbart, 1982). Rothbart

(1982) suggests that identifying a child as difficult may

both raise the expectations of problems where none have been

identified and suggest the existence of risk and

developmental continuity when in reality temperament

characteristics are not equally difficult at different










stages. She suggests that it may be more beneficial to

indicate that there are both positive and negative features

associated with each temperament characteristic.

Cameron and Rice (1986) have responded to Rothbart's

cautions by indicating that those working in pediatric

settings are invariably called upon to offer handling advice

in areas influenced by temperament characteristics. As

such, they maintain that the question is not whether to

provide this information, but rather how to most effectively

do so. Additionally, it should be noted, that the use of

descriptive rather than evaluative terms has also been

addressed by advocates of anticipatory guidance (Carey,

1982). Nonetheless, the issue remains as to whether it is

advisable to provide temperament screening and guidance as a

matter of course as compared to the alternative of utilizing

temperament data as part of an assessment when problems

arise. To address this issue it will be necessary for

researchers to more fully explore the potential benefits and

risks associated with this approach.

This review of the relevant literature suggests that

while temperament-based parent training programs are being

increasingly advocated as a means of facilitating less

stressful and more effective parenting, the work conducted

to date does not provide adequate support for the utility of

this approach. In particular, the failure to utilize

standardized outcome measures to assess changes in parental

adjustment, parent-child interaction, or child behavior










problems as well as the absence of control groups limits the

conclusions that can be drawn from the available literature.

Specific Aims

Given both the potential utility of temperament-based

parent training and the absence of adequate research support

for such programs, the goal of the present study was to

build on the existing clinical and theoretical base by

empirically examining the efficacy of a temperament-based

group parent training program. In particular, this

investigation focused on evaluating the effectiveness of the

program in alleviating maternal/familial difficulties and

reducing the level of child behavior problems.

It was hypothesized that relative to a wait-list

control group, subjects in the training group would

demonstrate reductions in anxiety and depression, as well as

increased comfort in the parenting role which involves

perceived parenting competence, feelings of being restricted

by the parenting role, satisfaction with the parent-child

relationship, ability to understand the child's needs, and

increased feelings of attachment to the child.

Additionally, improvement in child and family functioning

including spousal relationships, changes in family life-

style, and child behavior problems were also anticipated.

Based on results obtained by Webster-Stratton and Eyberg

(1982), however, significant changes in ratings of child

temperament were not expected at posttreatment.










Secondary to the primary objective, a subgoal in this

project involved more clearly relating the expected changes

in adjustment to participation in the parent-training

program. It was hypothesized in this regard, that the

improved adjustment expected in the active condition would

be accompanied by an increase in knowledge about child

temperament. It was expected, moreover, that attendance at

weekly meetings and completion of homework assignments would

relate to successful outcomes.














METHODS

Subjects

Demographic Data

The sample was comprised of 40 mothers with preschool-

age children. Two-thirds of the sample were recruited in

Portland, OR, and the remaining subjects were recruited in

Gainesville, FL. With two exceptions, all of the subjects

were white. The target children ranged in age from 2-years

9-months to 5-years 6-months (mean age 4). Sixty percent of

the children were male.

The majority of subjects were either married to or

living with the child's father (87.5%). In these families,

88% of the fathers, including one step-father, participated

in some portion of each assessment. In one family, the

maternal grandmother served as a secondary caretaker and

participated in the assessments. Additional demographic

information is provided in Table 1.

Criteria for Participation

Criteria for participation in the study were difficult

child temperament and the presence of reported

maternal/familial difficulties related to parenting. For

the purpose of subject selection, difficult child

temperament was defined by the following temperament

characteristics: negative mood, high intensity, low










Table 1

Demographic Information


Variable Mean Range



Mothers

Age 34 18-43

Education 15.5 12-18



Fathers

Age 35 22-44

Education 16 12-22



Family Data

Income 41 7-100

No of Children 2 1-5



Note: The values in the top two sections of the table
represent years. The values for income represent thousands
of dollars.










rhythmicity, low adaptability, high activity level,

withdrawal to new situations, and either high or low

distractibility. As discussed earlier, these variables have

been shown to be predictive of child behavior problems as

well as maternal and familial difficulties. Subjects

included in the sample rated their children as demonstrating

a minimum of three of these temperament characteristics as

demonstrated by a score which was at least one standard

deviation above the mean of an earlier nonclinic Gainesville

sample. As both low and high distractibility have been

found to be predictive of maternal and child difficulties, a

score one standard deviation from the mean in either

direction was accepted. The flexibility in the inclusion

criteria is thought to be in line with the goodness of fit

model in recognizing that various child temperament

characteristics may be predictive of difficulties within

different families (Buss & Plomin, 1985; Turecki & Tonner,

1985).

As activity level was one of the possible inclusion

criteria, an additional screening was considered necessary

in order to rule out the presence of Attention Deficit

Hyperactivity Disorder. Toward this end, potential subjects

were questioned as to whether medications had at any time

been tried to manage the target child's activity level or

behavior. While the plan was to exclude any subjects who

reported that this had been the case, this event never

occurred in this sample.










In addition to the above-mentioned child temperament

characteristics, several maternal/familial variables were

also chosen as inclusion criteria. These included Parenting

Stress Index (Abidin, 1983) attachment, perceived

competence, and restriction imposed by parenting role scales

as well as a modified version of the Impact on Family Scale

(Stein & Riessman, 1980). These variables were chosen from

a wider range of maternal psychosocial problems previously

shown to be associated with having a difficult child

(Sheeber & Johnson, 1991a) because they appeared most

directly related to problems of parenting. Mothers included

in the sample demonstrated difficulties on a minimum of two

of the four scales as evidenced by a score at or above the

75th percentile of the normative sample for the PSI

subscales and of an earlier nonclinic Gainesville sample in

the case of IFS. The 75th percentile was reported to be the

upper limit of the normal range for the PSI subscales

(Abidin, 1981). This level was chosen, rather than the 90%

cutoff considered to be indicative of severe stress because

it seemed more appropriate given a nonclinic population.

Additionally, the more conservative limit was considered too

restrictive to be utilized in conjunction with inclusion

criteria for child temperament.

Potential subjects were also screened to determine that

neither the mother nor the target child were presently

receiving psychological or psychiatric treatment.

Additionally, potential subjects who reported that their










children either demonstrated extreme behavioral disturbances

or were victims of severe trauma were excluded. As

described in more detail in the procedures section, this

information was obtained during a phone contact between the

potential subject and the investigator.

Measures

Demographic Questionnaire

A demographic questionnaire was completed by each

mother. The following information was gathered: age and sex

of target child, mother's age, race, marital status, number

of years of education completed by each parent, family

income, parents' occupations, the ages of other children in

the family, and whether there is a second caretaker in the

home.

Parent Temperament Questionnaire (PTO)

The PTQ is a parent-completed questionnaire developed

by Thomas, Chess and Korn (Thomas & Chess, 1977) for

assessing temperament in 3- to 7-year-old children. It is

comprised of 72 items, with eight items tapping each of the

nine temperament categories. Item content, derived from

parent interview protocols, was phrased so as to reflect

behaviors rather than attitudes (i.e., "my child selects..."

rather than "my child prefers..."). The relative frequency

of the specified behavior is marked on a 7-point scale.

Half of the items are phrased in terms of one extreme (i.e.,

high activity) and half in terms of the other extreme (i.e.,

low activity). The questionnaire was normed on a sample of










148 mothers with children between the ages of 3 and 7. No

significant differences were found as a function of the age

or sex of the child.

The PTQ has been utilized in numerous studies with both

normal and special populations (Black, Gasparrini & Nelson,

1981; Earls, 1981; Johnson, Basham, & Gordon, 1982; Sheeber

& Johnson, 1987; Thomas & Chess, 1977, 1984). Results from

this combined body of research indicate that the temperament

dimensions measured by this questionnaire are related to

child psychopathology, maternal adjustment, and the nature

of parent-child interactions; these data attest to the

suitabilty of this measure for child temperament research.

The PTQ has also demonstrated reasonable reliability

for most dimensions. One study (Katz-Newman & Johnson,

1986) reported retest reliabilities ranging form .48

(Distractibility) through .92 (Approach/Withdrawal) over a

2-week period. Internal consistency reliabilities (alpha

coefficients) ranging from .56 (Distractibility) to .72

(Approach/Withdrawal) were obtained for a sample of 77

children (Sheeber, 1987).

For the purposes of the present study, the individual

items were summed in their scaled directions and a mean

calculated to attain scores on each of the nine temperament

categories. Scores were computed such that a high score was

reflective of a relatively difficult behavioral style.










State/Trait Anxiety Inventory (STAI)

The STAI is comprised of two separate, 20-item self-

report scales designed to assess anxiety as both a state and

a trait (Spielberger, Gorusch, & Lushene, 1970).

Respondents are instructed to mark on a 4-point scale the

degree to which each item reflects how they feel at a

particular moment (State scale) and in general (Trait

scale). As previous research (Sheeber & Johnson, 1991a)

indicated that the two scales were not differentially

sensitive to the anxiety associated with having a

temperamentally difficult child, only the Trait scale will

be utilized in the present study.

The Trait scale demonstrates a relatively high degree

of internal consistency (.83-92) and retest (.73-.86)

reliability. The STAI has an extensive validation history

having been widely used in research on anxiety (e.g.,

Kendrick, Craig, Lawson, & Davidson, 1982; Rosenbaum &

Palmer, 1984). A number of studies have used the STAI to

measure maternal anxiety (Blumberg, 1980; Schubert-Walker &

Walker, 1980).

Parenting Stress Inventory (PSI)

The PSI (Abidin, 1983) is a self-report measure

designed to assess stress in the parent-child system which

may emerge as a function of parent characteristics, child

characteristics, and/or situational variables. It is

comprised of both maternal and child domains which reflect

stress as a function of maternal and child characteristics










respectively. Each domain consists of several subscales

which tap specific aspects of the domain. Additionally,

there is .a brief life events survey which taps stress

associated with situational variables. The final version of

the PSI was normed on a sample of 534 mothers of preschool

children seen in a pediatric clinic for both behavioral and

health problems. It is comprised of 101 items rated on a 5-

point scale. A higher score is indicative of greater

difficulty.

Several of the maternal domain scores will be used in

this study as measures of maternal adjustment. The parent

depression scale consists of items which tap clinical

depression as well as dissatisfaction with self or life

circumstances which might not be associated with a clinical

depression. The parent attachment scale reflects either the

absence of emotional closeness to the child or the parent's

perceived inability to accurately assess the child's

emotions and needs. Feelings of being dominated by the

demands 6f child-rearing at the expense of their own needs

for freedom and identity are reflected in the restrictions

imposed by parental role scale. The parents sense of

competence scale taps insecurity in the parenting role as

well as the absence of the expected reinforcement from

parenting. Finally, the relationship with spouse scale

reflects the absence of emotional and active support from

the other parent in the area of child management.










These scales have previously been found to be

associated with having a temperamentally difficult child

(Sheeber & Johnson, 1991a). Taken together with Casey's

(Cited in Abidin, 1983) finding that maternal domain scores

were significantly correlated (r=.40) with the Achenbach

Child Behavior Checklist (Achenbach & Edelbrock, 1983) these

results suggest that the maternal domain scores are

sensitive to maternal stress associated with having

difficult children. Additionally, the PSI has demonstrated

validity for research on maternal adjustment. The maternal

domain was found to be significantly correlated with both

the A-State (.71) and A-trait (.84) scales of the

State/Trait Anxiety Scale (Jenkins, cited in Abidin, 1983;

Lafiosca, cited in Abidin, 1983). Scores on the maternal

domain have, moreover, been shown to discriminate parents

who received perinatal coaching from those who did not

(Bristor, cited in Abidin, 1983), abusive form nonabusive

parents (Mash & Johnston, 1983) and mothers receiving

adequate support form their husbands from those who were not

(Lawrence, cited in Abidin, 1983). Factor analysis

supported the scale construction (Abidin, 1983).

The PSI also demonstrates adequate reliability. The

internal reliability coefficients for the maternal domain

score is .93, ranging from .55 (attachment) to .80

(depression) for the individual subscales. Moderate to high

test-retest reliability (.65-.91) has been demonstrated for

the maternal domain in four separate studies (Abidin, 1983).










Impact on Family Scale

The Impact on Family Scale was originally developed by

Stein and Riessman (1980) to assess changes in the normative

behavior of the family as a function of a child's chronic

illness. This scale was modified by the present

investigator (Sheeber & Johnson, 1991b) to reflect the

impact of behaviorally difficult children on the day to day

functioning of their families. The resulting questionnaire

contains 23 items judged to be reflective of family

disruption by 15 clinical child psychologists and graduate

students.

Preliminary construct validity was suggested by two

studies indicating that scores on the IFS were significantly

higher for conduct disordered children than for children not

demonstrating behavior problems as well as for children

demonstrating difficult as compared to easy temperament

(Sheeber & Johnson, 1991a). The IFS also demonstrated high

reliability. The split-half correlation coefficient

(Spearmar-Brown Formula) for odd and even items was .96.

Mean item to total correlation (Chronbach's Alpha) was .93.

Additionally, retest reliability was .96 across a mean time

span of 4 days (range 1-14 days) between the time of initial

contact and the first clinic visit. As these

administrations were conducted over a short period of time

(range 1-14 days; mean time span 4 days) it is felt that

these scores do not provide conclusive findings regarding

the stability of the measure. However, as the scores










demonstrated only slight changes in both directions at the

second administration, they do indicate that the IFS is not

subject to a reduction in the severity of reported

complaints between the time of initial contact and the first

clinic appointment. A copy of the modified IFS is provided

in Appendix A.

Cleminshaw-Guidubaldi Parent Satisfaction Scale

This self-report measure is comprised of five

empirically derived subscales reflecting various dimensions

of parenting satisfaction: spouse support, child-parent

relationship, parent performance, family discipline and

control, and general satisfaction (Guidubaldi & Cleminshaw,

1985). Item content was derived from open ended interviews

with parents as well as a review of relevant literature.

Items are scored on a 4-point scale and are balanced for

directionality. The original sample included 52 fathers and

78 mothers.

While each of the above mentioned scales taps

information likely related to parenting a temperamentally

difficult child, there is substantial overlap with the PSI

dimensions. For this reason, only the two scales providing

information not assessed by the PSI will be utilized: child

parent relationship and family discipline and control. The

former reflects the parents' satisfaction with their own

relationship to the child, while the latter taps

satisfaction with the spouse's parenting particularly as

regards disciplinary matters. For the purpose of the










present study, mothers completed the child parent

relationship scale. Fathers, or in some cases secondary

caretakers, completed the family discipline scale. The

family discipline scale was included to provide an

additional source of information regarding the mother-child

relationship.

High internal consistency reliabilities (Chronbach's

alpha) have been demonstrated for these scales (parent child

relationship, r=.86; family discipline r=.82). Preliminary

construct validity was demonstrated by moderate correlations

with measures of marital and general life satisfaction.

Additionally, the parent-child relationship scale has been

shown to discriminate between clinic and nonclinic mothers

(Mouton & Tuma, 1988). As this is a relatively new measure

additional psychometric data is not presently available.

Findings based on this measure would necessarily be

considered preliminary in nature. Nevertheless, due to the

absence of other non-observational measures for assessing

parent-child relationships, it is felt that the inclusion of

this instrument would be a worthwhile addition to the

proposed research.

Child Behavior Checklist (CBCL)

The CBCL (Achenbach & Edelbrock, 1983) is a parent-

completed measure designed to assess social competence and

behavior problems in children between the ages of 4 and 16.

Normative data are based on a sample of 2300 clinic and 1300

nonclinic children. The Behavior Problem Scale consists of










118 items rated on a three-point scale; a higher score is

indicative of greater difficulty. Separate profiles are

used for each sex and for three age ranges. These profiles

depict the child's behavioral pattern across a range of

empirically determined categories which reflect both broad-

band (internalizing, externalizing) and narrow-band

syndromes (e.g., depression). As the sample for the present

study was not large enough to analyze results separately by

age group and as the profiles were not used for diagnostic

purposes, the 4-16 year old version was used for all

subjects rather than using the 2-3 year old version with

younger subjects; this approach was supported by the author

of the measure (T. M. Achenbach, personal communication,

February 1989).

The CBCL demonstrates adequate test-retest

reliabilities over a one-week (r=.89) and six-month (r=.60s)

periods. Additionally, between parent agreement is

moderately high (r=.66; Beck, 1987). The CBCL has been

utilized extensively in research on child psychopathology

and treatment. Profiles scores have been shown to be

sensitive to treatment changes arising from parent-training

interventions with conduct disordered children (Webster-

Stratton, 1984). Additionally, behavioral disturbances on

this measure were found to be related to difficult child

temperament in a sample of 6- to 7-year old children

(Garrison, Earls, & Kindlon, 1984).










The broad-band factors were utilized to assess changes

occurring as a function of the intervention. The CBCL was

completed independently by mothers and when available,

fathers or secondary caretakers.

Parent Daily Report (PDR)

Originally developed by Patterson, Reid, Jones, and

Conger (1975), this parent observation measure of child

behavior problems was designed to alleviate the major

difficulties associated with parent report and behavioral

observation measures. More time and cost effective than

independent behavioral observations, the PDR was aimed at

providing a more valid and reliable approach to obtaining

parent information data. The more recent version

(Chamberlain & Reid, 1987) consists of 34 items: 33 are

deviant child behaviors and one taps parental use of

physical discipline. It is conducted during brief (5-10

minute) phone interviews. Parents are asked to indicate the

occurrence or nonoccurrence of each event (total behavior

score) or targeted events (target behavior score) during the

preceding day. Targeted events represent a select group of

child behaviors which the parent has previously identified

as being in the child's behavioral repertoire. The score

represents a mean frequency of symptoms per day. Several of

the behaviors tapped by this measure have been reported to

be evidenced by temperamentally difficult children (e.g.,

temper tantrums, irritableness, fearfulness). Normative










data for this version are based on 81 children between four

and ten years of age.

Though the measure yields a score for both total

deviant behaviors and targeted behaviors the latter

demonstrates more adequate psychometric characteristics and

was consequently utilized in this study (Chamberlain & Reid,

1987). Test-retest reliability for the target behavior

scale was .82 over a 4-week period in a nonclinic sample.

One study with a clinic population found no differences in

weekly target scores across a 4-week baseline. Though PDR

scores may be inflated on the first day, they are stable

thereafter; the suggested procedure of conducting three

interviews per assessment period controls for this potential

difficulty. Interparent agreement ranged from .51 (clinic)

to .89 (nonclinic). Mean entry to entry intercaller

agreement ranged from .85 (clinic) to .98 (nonclinic).

The PDR has demonstrated significant concurrent

validity with the Family Interaction Coding System (r=.69;

Patterson, 1982). This level of agreement is considered to

be exceptionally high given the different time span and

exact behaviors which are tapped by the two measures.

McMahon and Forehand (1988) report that the PDR has been

used extensively as a measure of treatment outcome. While

Patterson (1982) cautions that the PDR is reactive relative

to other observation procedures, particularly as regards the

influence of maternal mood, steps were taken in this study

to reduce the influence of this potential problem. In










particular, three separate interviews were conducted per

assessment period and both maternal and paternal reports of

child behavior were obtained.

For the purposes of the present study, it is felt that

the information provided by the PDR is complementary to

that provided by the CBCL due to differences in two

important domains. Firstly, the data from the PDR targets

the specific behaviors identified by parents as problematic.

Secondly, it taps information from parent observation for

the immediately preceding 24 hours, thereby increasing the

likelihood of accurate reporting.

Parent Test of Temperament Knowledge (PTTK)

This measure was developed by the investigator for use

as a manipulation check in this project. The purpose of the

instrument was to assess whether subjects had an increased

understanding, subsequent to the intervention, of the role

of child temperament in both understanding child behavior

and developing effective behavior management approaches.

The 'first step in the development of the measure was

generating items to tap the information covered in each

session. Approximately five items were developed for each

session. Given the PTTK's sole use as a measure of acquired

information, it was felt that the primary validity issue

pertained to the adequacy of the items in sampling the

content of the training program. To address this, 14

psychology faculty members and students familiar with the

area of child temperament were asked to review the measure









in conjunction with the training manual. They provided

feedback regarding the following: clarity of the items,

adequacy of the items in sampling the domain of material,

and difficulty of the items. Based on the feedback received

individual items were clarified and the measure was

shortened to 23 items.

The next step in the development of this measure was to

administer the 23 item PTTK to a sample of 15 mothers whose

children attended an on-campus preschool. These women

completed the instrument on two occasions approximately

three weeks apart. Based on the data they provided, items

were eliminated which were answered correctly by such a

large percentage of subjects that their ability to reflect

increased knowledge as a function of treatment was expected

to be extremely limited. Nine items which were answered

correctly by at least 87% of this sample were so eliminated.

The resulting 14 item test had a test-retest reliability of

.71 (R < .003). A copy of this measure is provided in

Appendix 'A.

Procedure

Recruitment and intake assessment

Subjects were recruited via fliers distributed to area

preschools as well as advertisements placed in local

publications targeting families with young children. The

fliers and advertisements indicated that a parenting group

was available for mothers of temperamentally difficult

children as part of a research project. Several temperament










characteristics such as "moody", "active", "dislikes

changes" were listed in order to help parents identify the

appropriateness of the program for their needs. The initial

contact with subjects consisted of a parent-initiated phone

call during which a brief description of the treatment and

research program was offered, and a series of screening

questions were conducted to determine the following: whether

temperament characteristics appeared to be primarily

contributing to parenting difficulties; whether the child's

behavior was suggestive of psychopathology; whether the

child's behavior appeared to be largely in response to a

recent or severe trauma. Additionally other questions

addressed exclusion criteria: current psychological/

psychiatric treatment of mother or child or previous trials

of medication to control hyperactivity or behavior.

Potential subjects who were either not interested in the

services offered or who did not seem appropriate based on

this brief screening were offered referrals to service

providers in the community.

The next level of screening involved an appointment

between potential subjects and the investigator. At this

meeting subjects signed an informed consent form which

provided information regarding the objectives and procedures

for the research as well as the potential risks and benefits

of participating. Subjects had an opportunity to ask

questions about the research and were informed that the










investigator could be contacted if questions or concerns

arose at a later time.

Subsequent to the informed consent procedures, subjects

were administered a series of questionnaires which tapped

the aforementioned inclusion criteria. The measures

administered were Parent Temperament Questionnaire,

Demographic Questionnaire, Parenting Stress Index, Maternal

Domain, and the Impact on Family Scale-Revised. Again at

this point, subjects who either did not meet inclusion

criteria or who decided not to participate were made aware

of appropriate community resources. Subjects who were

appropriate and interested in continuing were mailed the

remainder of the questionnaires and contacted by phone for

the PDR interviews prior to the first group meeting.

Fathers and secondary caretakers were then provided with

informed consent forms, mailed questionnaires, and contacted

both for PDR interviews and so as to address any questions

or concerns that they had about the program.

Group assignment

Subjects were randomly assigned to either the parent-

training group or to a wait-list control group. Subjects

were informed that those assigned to the wait-list condition

would be placed in a parent-training group four months

later; this information was provided both verbally and in

the written informed consent form. Initially 20 subjects

were assigned to the wait-list condition and 21 to the

parent-training group; however, one subject in the parent-










training group discontinued before the end of the program.

Subjects assigned to the parent-training condition

participated in three separate groups of between 6 and 8

members each. Groups were divided by location: Gainesville,

SW Portland, and SE Portland. In addition to making the

groups more accessible to subjects, it was felt that the

smaller group size allowed more time for individual

questions and discussions. This same procedure was

maintained when the program was conducted with the wait-list

group.

Parent-training program

The model of child temperament used in this program

most closely approximated that developed by Thomas and Chess

(1977). A benefit of this framework over alternative models

was that it facilitated building on the investigator's

previous work regarding the relationship between child

temperament and maternal adjustment (Sheeber & Johnson,

1991a) as well as on the temperament-focused program

developed by Turecki (Turecki & Tonner, 1985) as discussed

below. Additionally, the greater breadth of dimensions used

in the Thomas and Chess model was felt to be advantageous

from a clinical perspective in that it facilitated

addressing a wider range potential "fit" issues.

Based on the goodness-of-fit notion, the protocol for

the parent-training program was geared toward utilizing an

understanding of child temperament to enhance the match

between child characteristics and parental demands. The










program was largely based on the framework described by

Turecki (Turecki & Tonner, 1985), though the approach was

modified to include recommendations offered in the works of

Thomas and Chess (1986;1987), to reflect the investigator's

understanding of child temperament and behavior management,

and to facilitate the use of the program in a group format.

The groups, led by the investigator, met weekly for 1

1/2 2 hour sessions over the course of nine weeks. The

initial focus of the program was on familiarizing parents

with the nature of child temperament, describing its role in

understanding child behavior, and exploring the parental

feelings and family difficulties related to parenting a

difficult child. Subsequent to this, the program focused on

techniques for managing temperament-related behavior

problems. General information on behavior management was

incorporated into this phase of the program. Homework was

assigned to facilitate the use of material discussed in

sessions.

In order to verify that the protocol for the program

was followed accurately a checklist was developed that

detailed the objectives for each session. This checklist

provided a guide to the investigator in conducting the

sessions. Additionally adherence to the protocol was rated

by an undergraduate research assistant who observed during

each session. The average protocol compliance across the

three groups was 96%. Slightly more than 25% of these

sessions were rated by a second undergraduate observer to










assess reliability. The correlation between the two raters

was .92 (R <.05).

Posttreatment assessment

The final session of the program was used to administer

the posttreatment assessment questionnaires. With the

exception of the form for acquiring demographic data, all of

the measures administered in the two-stage intake assessment

were completed at this session. Subjects in the wait-list

group completed these questionnaires by mail during the same

week. The PDR interviews for both groups began after this

final session.

Provision of training program for wait-list subjects

Subjects in the wait-list condition began the parent-

training program two months after the posttreatment

assessment. The groups for wait-list subjects at the two

Oregon locations were led by the investigator. The group

for wait-list subjects in Florida was led by another

doctoral candidate who had been trained in the program.

Fourteen of the twenty wait-list subjects chose to

participate in the program at this point. However, there

was substantial attrition and only 7 of the subjects were

involved by the end of the eight weeks.














RESULTS

Homoqeneity of Sample

As discussed earlier, the sample for this study was

drawn from both Portland and Gainesville. Given this, the

first step was to determine whether any significant

differences existed between subjects as a function of

location which would preclude they're being considered one

sample. Initially, the groups were compared for similarity

as regards demographic variables. Chi Square analyses

indicated that both sex of the child, X2(1, N=39) = .3, ns

and marital status, X2(3, N=39) = 1.39, ns were distributed

evenly across locations. A manova conducted on the remaining

demographic variables similarly indicated no significant

difference as a function of location, F(6,30) = 1.68, ns.

This was also the case for a manova conducted on the

temperament characteristic scores as assessed at

pretreatment, F(9,29) = .66, ns.

A series of manovas were conducted to determine whether

differences existed on the dependent variables at

pretreatment as a function of location. The choice of

variables to be entered into each model were based on both

rationally defined content areas and the number of

observations available for a given variable. As indicated

below, these analyses were also nonsignificant: variables

related to parenting (F(4,35) = 2.07, ns; competence,










restriction, competence, parent-child relationship);

maternal affect (F(2,36) = .02, ns; depression, Anxiety);

child behavior as rated by mother (F(2,37) = .57, ns; CBCL-

internalizing and externalizing); child behavior and

maternal discipline approaches as rated by father (F(3,26) =

1.23, ns; CBCL-internalizing and externalizing scales,

family discipline); and family functioning (F(2,34) = .17,

ns.; IFS, relationship with spouse). Due to reduced sample

size, PDR scores were analyzed using t-tests rather than

including them in the manova models. No significant

differences between locations emerged for mothers, t(32) =

3.03, ns or fathers, t(21) = .01, ns. Finally, maternal

knowledge about temperament as assessed by the PTTK was

equivalent across the two sites, t(34) = 2.04, ns. Based on

these results, data from subjects across the two locations

were analyzed jointly.

Equivalence of Subjects across Conditions

As reported in the procedures section, assignment to

either the parent-training or the wait-list condition was

random in order to increase the likelihood that the groups

would be comparable on relevant factors. However, as

randomization is less effective with relatively small sample

sizes, it was considered important to assess the degree to

which equivalence of groups was achieved.

The procedure for doing this was essentially the same

as that discussed above for determining the homogeneity of

the sample across locations. Initially, the groups were










compared for similarity as regards demographic variables.

Chi Square analyses indicated that there were no significant

differences between the two groups as regards sex of the

target child X2(1, N=39) = .82, ns, marital status X2(3,

N=39) = 4.18, ns or sibling order X2(1, N=39) = 1.38, ns. A

manova conducted on the remaining demographic variables

similarly indicated no significant difference as a function

of condition F(6,30) = .26, ns. This was also the case for

a manova conducted on the temperament characteristic scores

as assessed at pretreatment F(9,29) = 1.06, ns.

A series of manovas were then conducted to determine

whether the groups differed on the dependent variables at

pretreatment. The variables were entered into the analyses

using the same groupings as described above. The results of

these analyses were as follows: variables related to

parenting (F(4,35) = 1.24, ns; attachment, restriction,

competence, parent-child relationship); child behavior as

rated by mother (F(2,37) = .53, ns; CBCL-internalizing and

externalizing); child behavior and maternal discipline

approaches as rated by father (F(3,26) = .47, ns; CBCL-

internalizing and externalizing scales, family discipline);

family functioning (F(2,34) = .65, ns.; IFS, relationship

with spouse) and maternal affect (F(2,36) = 3.13, R <.06.;

depression, Anxiety). These analyses indicated that the

groups were not significantly different at pretreatment

although the maternal affect variables approached

significance. As shown in Table 2, the treatment groups










scores were slightly higher on these two variables. Due to

reduced sample size, PDR scores were analyzed using t-tests

rather than including them in the manova models. No

significant differences between groups emerged for mothers

t(32) = .01, ns, or fathers t(21) = .26, ns. Finally,

maternal knowledge about temperament as assessed by the PTTK

was equivalent for the two conditions t(34) = 1.09, ns.

Taken together, these results suggest that random

assignment was effective in producing two comparable groups.

Mean scores on each of the dependent variables are presented

by condition in Tables 2-4

Program Effectiveness

Groups Comparisons

The primary goal of the study was to assess the

effectiveness of the parent-training program in alleviating

maternal/familial difficulties and reducing the level of

child behavior problems. In order to address this goal, a

series of mancovas were conducted to compare group scores at

posttreatment; pre-scores were entered as covariates in

these analyses. As in the previous analyses, the choice of

variables to be entered into each model was based on both

rationally defined content areas and the number of

observations available for a given variable. The number of

observations varied depending on the source (mother vs

father) and the format of the measure (questionnaire vs

interview); additionally, small amounts of data were lost








Table 2

Mean Scores for Parenting and Maternal Affect Variables as a
Function of Time and Condition

Time

Parenting
Variables Pre Post


Parent-training condition

Parent-child 22.19 ( 3.03) 18.38 ( 2.75)
relationship

Attachment 15.33 ( 3.50) 13.62 ( 2.36)

Competence 38.19 ( 4.64) 32.95 ( 4.74)

Restriction 24.19 ( 4.96) 20.76 ( 4.57)

Depression 28.19 ( 5.56) 20.52 ( 4.11)

Anxiety 45.57 ( 8.24) 38.30 ( 7.37)

Wait-list condition

Parent-child 20.89 ( 4.58) 20.84 ( 4.18)
relationship

Attachment 15.16 ( 2.87) 15.32 ( 3.18)

Competence 34.84 ( 6.98) 36.47 ( 6.20)

Restriction 24.68 ( 5.35) 23.05 ( 5.70)

Depression 24.00 ( 5.25) 24.05 ( 6.22)

Anxiety 43.44 (10.43 43.44 ( 9.51)


Note. The values in parentheses are standard deviations.









Table 3

Mean Scores for Family Functioning Variables by Time and
Condition

Time

Family
Variables Pre Post


Parent-training condition

Impact on Family 46.94 (12.41) 31.00 (14.85)

Relationship with 22.00 ( 5.19) 21.72 ( 4.39)
spouse

Family Discipline and 14.06 (1.69) 12.53 (2.06)
control


Wait-list condition

Impact on Family 42.95 (14.02) 38.68 (16.56)

Relationship with 21.79 ( 4.60) 22.58 ( 6.10)
spouse

Family Discipline and 14.21 (3.26) 13.25 (3.31)
control


Note. The values in parentheses are standard deviations.








Table 4

Mean Scores for Child Behavior Variables by Time and
Condition


Time

Child Behavior
Variables Pre Post


Parent-training condition

Mother Completed

CBCL-externalizing 66.52 (7.47) 58.24 (9.60)

CBCL-internalizing 61.43 (9.38) 55.09 ( 9.76)

Parent daily report 11.93 (3.50) 8.28 (2.96)

Father Completed

CBCL-externalizing 61.17 (7.64) 58.50 (9.24)

CBCL-internalizing 58.71 (6.81) 56.56 (9.13)

Parent daily report 10.59 (4.52) 7.85 (4.14)


Wait-list condition

Mother Completed

CBCL-externalizing 63.37 (10.33) 63.06 ( 9.38)

CBCL-internalizing 60.68 (10.00) 61.28 ( 8.25)

Parent daily report 11.78 ( 4.52) 10.79 ( 5.11)

Father Completed

CBCL-externalizing 57.71 (15.44 57.58 (13.50)

CBCL-internalizing 54.43 (12.93) 55.17 (10.56)

Parent daily report 9.52 ( 5.59) 8.46 ( 4.56)


Note. The values in parentheses are standard deviations.










due to error (e.g., participants completing the wrong

items). Significant mancovas were followed up with ancovas

to identify the specific variables differentiating the

groups at posttreatment.

The first set of variables entered related to maternal

comfort in the parenting role. This analysis indicated that

significant differences emerged as a function of condition

F(4,31) = 4.61, p<.01. As shown in Table 5, subsequent

univariate analyses indicated changes in perceived parenting

competence, satisfaction with the parent/child relationship,

and attachment to the child. As assessed by this measure,

attachment reflects both a sense of closeness to the child

and the belief that one understands the child's needs. It

is notable, that feelings of being restricted by the demands

of parenting were not significantly influenced by group

participation.

The impact of the training program on maternal affect

was demonstrated in the second mancova F(2,32) = 11.44,

p<.001. The univariate analyses in Table 5, revealed that

both anxiety and depression differentiated the groups at

posttreatment. Family variables were entered into the next

analysis which also emerged as significant F(2,32) = 5.92,

2<.001. Subsequent univariate analyses indicated a

significant difference between the conditions on a modified

version of the Impact on Family Scale (IFS) which taps

negative changes in life-style as a function of child

behavior. Relationship with spouse, which as assessed by










Table 5

Univariate Group Differences at Posttreatment


Variables


F value (df)


Parenting

Parent-child relationship

Attachment

Competence

Restriction

Maternal affect

Depression

Anxiety

Family functioning

Impact on family

Relationship with spouse

Mother-rated child behavior

CBCL-externalizing

CBCL-internalizing

Parent daily report


11.09

4.79

11.06

2.72



14.63

15.13


12.20

.39


9.33

6.75

6.86


(1,34)

(1,34)

(1,34)

(1,34)



(1,33)

(1,33)



(1,33)

(1,33)



(1,35)

(1,35)

(1,28)


R<.05 *E<.01


***<.001
E<.001










this scale measures emotional and instrumental support in

the area of child management, did not differentiate the

groups. These univariate analyses are presented in Table 5.

The remaining analyses assessed differential change in

child behavior as a function of condition. As rated by

mothers, significant differences existed at posttreatment on

both CBCL data F(2,34) = 4.72, R<.05 and the Parent Daily

Report F(1,28) =6.86, R<.01. The univariate analyses

presented in Table 5 indicate that differences emerged for

both internalizing and externalizing behaviors. Notable,

however, is that these same variables rated by fathers did

not emerge as significant: PDR F(1,20) =1.98, ns; CBCL and

family discipline F(3,19) = .78, ns. The family discipline

scale entered into this analysis assesses the father's

satisfaction with the mother's parenting skills.

Tables 2-4 provide means for each of the dependent

variables at pre- and posttreatment. An examination of the

means by condition indicate that all of the observed

differences at posttreatment were in the expected direction.

That is, the differences observed between the conditions at

posttreatment are attributable to lowered difficulty ratings

for subjects in the parent-training condition.

It is also notable in this context, that consistent

with expectations, no significant differences arose in

temperament ratings as a function of condition. The

temperament variables were entered into three mancovas. The

results were as follows: adaptability, approach/withdrawal,










rhythmicity F(3,33) = .52, ns; mood, intensity, activity

F(3,33) = 2.20, ns; persistence, distractibility, threshold

of responsiveness F(3,33) = 1.33, ns.

Manipulation Checks

Given the significant group differences which emerged

at posttreatment in the majority of maternal, family, and

child variables, the next step involved more clearly

relating these changes to participation in the parent-

training program. A paired t-test was conducted on the

Parent Test of Temperament Knowledge (PTTK) scores within

each condition to assess change from pre- to posttreatment.

The results indicated a significant difference in knowledge

from pre- to posttreatment for subjects in the parent-

training condition, t(21) = 8.49, R<.0001. No difference in

knowledge was observed for the subjects in the wait-list

condition, t(19) = 0, ns. These results suggest that the

improvement observed parent-training subjects was related to

information gained in the course of the program.

A more detailed level of analysis examined the

influence of attendance and homework completion on outcome

for subjects in the active condition. For each of the

variables which had differentiated the groups, a

corresponding variable was created by removing the variance

in the post score attributable to the pre-score. Spearman

correlations were then run between each of these newly

created variables and the homework and attendance variables.

These residual variables were created because this approach










fit conceptually with the main analyses in which pre-scores

were used as covariates. Spearman correlations were run

because the data were not normally distributed.

These analyses were hampered, however, by low

variability in both attendance and homework completion.

Reduced variability in attendance was attributable to the

almost uniformly high attendance rate among participants;

average attendance was seven of eight sessions (range 5-8).

As regards the homework data, it was initially proposed that

the homework variable would be based on daily ratings,

yielding a percentage of days in which skills were

practiced. However, difficulty in obtaining accurate

subject recordings led to the data being recorded in a

dichotomous fashion, i.e., did the subject practice in the

preceding week or not, thus substantially reducing expected

variability (mean = 7; range 5-8).

The results are presented in Table 6. As can be seen,

only a portion of these correlations was significant. As

would be expected, the correlations were negative,

indicating that increased participation was related to

reduced levels of difficulty. Based on these results,

homework completion appears to be related to improvements in

mother-reported child behavior and satisfaction with the

parent-child relationship as well as with fewer reported

negative life-style changes. Attendance was related to

attachment as defined above, and improved child behavior.










Table 6

Relationship Between Group Participation and Outcome



Spearman Correlations



Variables Homework Attendance



Parenting

Parent-child relationship -.33* -.01

Attachment .23 -.47

Competence .08 -.05

Maternal affect

Depression -.11 .15

Anxiety -.05 .14

Family functioning

Impact on family -.43 .22

Mother-rated child behavior

CBCL-externalizing -.38* -.05

CBCL-internalizing -.15 -.09

Parent daily report -.39** -.46*


*<.05


**<.
E<.01










Clinical Impact of Group Participation

While the findings presented thus far indicate that

the parent-training program produced statistically

significant changes across many domains, an additional

question related to the clinical significance of these

results. Although at the current time, there is not a

uniformly accepted method for determining the clinical

significance of obtained results (Kazdin, 1986), a

frequently used approach is to examine the extent to which

treatment brings the sample into the normal range (Jacobson

& Truax, 1991; Kazdin, Esveldt-Dawson, French, & Unis, 1987;

Kendall & Norton-Ford, 1982). The effectiveness of the

treatment in this regard can be examined on both group and

individual levels by using normative data as a reference

(Kazdin, 1986; Kendall & Norton-Ford, 1982). Normative data

are available for each of the measures used in this

investigation with the exception of the modified Impact on

Family Scale (Sheeber & Johnson, 1991b) and the subscales of

the Cleminshaw-Guidubaldi Parent Satisfaction Scale

(Guidubaldi & Cleminshaw, 1985). When a cut-off

differentiating the normal from clinic range was provided in

the manual, this value was used for comparisons. This was

the case for the Child Behavior Checklist (Achenbach &

Edelbrock, 1983) and the Parenting Stress Index (Abidin,

1983). Alternatively, a standard of one standard deviation

above the mean was used as a cut-off; this was the case for

the Parent Daily Report (Chamberlain & Reid, 1987) and the










STAI (Spielberger, Gorusch, & Lushene, 1970). Only

variables for which a statistically significant group

difference had emerged at posttreatment were evaluated with

regard to the clinical significance of the results.

Group means. In addressing the question of clinical

significance at the group level, the relevant question

pertains to whether at posttreatment, group means are within

the normal range. The cut-off for the normal range on the

Parenting Stress Index subscales is the 75 percentile

(Abidin, 1983). Mean scores for both the control and

treatment groups were outside the normal range at

pretreatment on each of the variables examined. At

posttreatment, mean scores for the treatment group were

within normal limits while those for the control group

continued to be outside of these limits. On the mother-

rated Parent Daily Report, subjects in both groups were

outside of normal limits at both pre- and posttreatment

assessments. These results are presented in Table 7. As

subjects in both groups were within normal limits at

pretreatment on mother-rated CBCL scales as well as on the

STAI, posttreatment scores were not examined with reference

to normative scores for these variables.

Individual cases. Although group means are useful in

providing a global index of treatment efficacy, they do not

reflect the progress of individual subjects. Tests of the

significance of difference between two proportions (Bruning

& Kintz, 1977) were performed to address this issue. In










particular, these tests were used to determine whether the

proportion of subjects falling within the normal range at

posttreatment was significantly different across the two

groups. Subjects were excluded from these analyses if their

pretreatment scores on a given variable were within the

normal range. This test was chosen over the more

conventional Chi-square, because in a Chi-square analysis

there would have been cells with too few observations to

provide valid tests.

The results indicated significant differences in the

proportions of subjects in the two groups who fell within

normal limits on the STAI (Z=3.06, p<.01) and the perceived

parenting competence (Z=4.5,E<..0001) subscale of the

Parenting Stress Index. Results of the remaining analyses

were nonsignificant (attachment Z=0; depression Z=1.47;

CBCL-internalizing scale Z=.83; CBCL-externalizing scale

Z=.59; PDR Z=0).














DISCUSSION

The results of this study provide preliminary support

for the efficacy of a temperament-focused parent-training

program. Mothers who participated in this program reported

increased comfort in the parenting role. In particular,

they reported greater satisfaction in their relationships

with their children, feeling more competent as parents, and

experiencing more attachment to their children as defined by

both feelings of emotional closeness and greater ease of

understanding the child's needs. Additionally, these

mothers indicated that they had experienced a more general

improvement in mood as regards reduced levels of anxiety and

depression.

Mothers also reported improvements in child behavior.

Mothers who participated in the program rated their children

as evidencing less problematic behavior in both the

internalizing and externalizing spectrums. Moreover, these

changes were observed by the mothers in the children's daily

behavior. Finally, these mothers indicated that the

negative impact of their children's behavior on their life-

style had been reduced. It is likely that this latter

change was a function of both improvements in the children's

behavior and increased maternal confidence in their own

ability to handle difficult child behavior when it arose.










Contrary to expectations, neither feelings of being

restricted by the parenting role or spousal relationships

were positively influenced by group participation. It had

been hypothesized that the greater ease in handling

difficult child behavior provided by the use of temperament-

based approaches would allow mothers to allocate more of

their resources to their own needs and interests. The

results suggest that the intervention may not have been

sufficiently powerful to achieve that end. Alternatively,

it is possible that spending two hours a week attending a

parenting group, in addition to the paperwork and homework

demands created by participation in this study, served to

offset that potential benefit. Future analysis of data from

a follow-up assessment will be helpful in teasing apart

these possibilities; if the demands of group participation

were operative in offsetting improvement in this area,

improvements should be evident at that later date.

On the other hand, with the benefit of hindsight, the

finding regarding mothers' relationships with their spouses

is not altogether surprising. The relationship with spouse

subscale of the PSI largely taps provision of emotional and

instrumental support in caring for the children. As fathers

did not participate in the program, gaps were created in

both the knowledge base and the parenting strategies likely

used by mothers and fathers thus making it difficult for

fathers to be in supportive roles. Clinical impressions

derived from leading these groups supports this conclusion.










Although encouraged to share information and ideas acquired

in the program with their partners, mothers routinely

reported feeling frustrated due to the difficulty of doing

so. In fact, the primary recommendation offered by group

participants for improving the program was to encourage

father participation.

Overall the results presented are consistent with the

goals and expectations for temperament-based parent training

as originally proposed by Thomas and Chess (1977) and more

recently supported by others interested in clinical

applications of child temperament data (Cameron & Rice,

1986; Carey, 1982). These results specifically support an

approach to temperament-based parent education which does

not rely on a pre-conceived definition of "difficulty".

This point is worth highlighting given both concerns about

the relative risks and utility of the "difficult" label

(Rothbart, 1982) and the "goodness of fit" concept which

emphasizes that the specific temperament characteristics

found to pose difficulties for the child or parent will vary

as a function of child's age, parental demands, and other

contextual factors (Thomas and Chess, 1977; Turecki and

Tonner, 1985). The flexibility of the inclusion criteria

and the process of recruiting subjects solely through self-

referral were both designed to take these concerns into

consideration.

The findings presented here extend previous work in a

variety of ways. Firstly, the use of validated outcome









measures and the collection of post-assessment data from all

study participants, provide for both greater confidence in

the results and greater clarity as regards their meaning.

The delineation of areas in which the training was and was

not successful facilitates future development of parent-

training programs in a way that global ratings of

satisfaction or change do not. Additionally, and perhaps

most importantly, the inclusion of a control group and the

findings of increased knowledge about temperament for

subjects in the active condition, increase the likelihood

that observed changes are reasonably attributable to program

participation. Finally, the finding that temperament-based

training can be provided effectively in a group format is

important given the greater cost-effectiveness of this

approach over individual sessions.

There are, however, aspects of the findings, which in

running counter to hypothesized outcomes have left questions

to be addressed. Primary among these was the absence of

expected improvements in father-rated indices of child

behavior and maternal parenting skill. With respect to the

former, it is considered likely and most parsimonious to

accept that this is an accurate reflection of child behavior

in the context of father-child interactions. As fathers

did not participate in the program, there is no reason to

expect that their style of interacting with the children was

altered in such a way as to promote either smoother

interactions or improved child behavior. Although, the










present study, having not utilized direct observations of

parent-child interactions, cannot speak to this directly,

this explanation is consistent with both the goodness of fit

model and with knowledge about the situational specificity

of child behavior.

Several potential explanations exist for the lack of

change in father ratings of maternal parenting skill. Two

related possibilities are that mothers either did not change

their parenting behaviors or that the changes made were not

congruent with fathers' views on appropriate discipline.

Again, the absence of observational measures precludes

direct assessment of these hypotheses at this time.

However, the improvements in mother-rated child behavior

suggest that mothers may have been handling child behavior

differently than in the past. This is especially the case

given the finding that homework completion, which typically

involved practicing new behavior management approaches, was

predictive of improvement in child behavior. Alternatively

it is feasible that fathers were not sensitive to the

changes that occurred over a relatively brief period of

time. In future analyses of data from the follow-up

assessment it will be interesting to note whether father

ratings of maternal skill change over time. Finally, as

discussed in the measures section, the Cleminshaw Guidubaldi

measure is relatively new and its psychometric qualities are

therefore less than certain. It is possible therefore, that

the particular scale used to obtain father ratings in this










area may not be as sensitive to treatment changes as was

originally hoped.

Another area in which the results ran counter to

hypotheses was the lack of consistent relationships between

degree of improvement and either homework completion or

attendance. The significant results which did emerge

provide additional support for the link between program

participation and observed improvement. Particularly strong

in this regard was the pattern of results suggesting a

relationship between homework completion and child behavior.

This is notable given that unlike variables more specific to

maternal functioning, child behavior would only be likely to

change if mothers were practicing their skills at home with

their children.

Nonetheless, as previously discussed, only a subset of

these analyses were significant. Several alternatives exist

to explain the inconsistency observed in this data.

Firstly, based on the design of this study, it is not

feasible to determine which components of the training

program were operative in producing observed changes or how

many sessions would be necessary to achieve these results.

It is possible, therefore, that the observed inconsistency

in these findings reflects the fact that fewer sessions than

are currently in the program would have been necessary to

obtain many of the desired improvements. Additionally, as

discussed in the results section, both good attendance and

methodological difficulties in collecting the homework data










contributed to limited variability in the data. It is

possible that this restricted range placed limits on

establishing relationships between these variables and

treatment outcome. Taken together, these data suggest that

within the narrow range observed in this study, somewhat

reduced participation may not effect outcome except in

regard to child behavior. It awaits future research to

determine the components of the program which are most

critical and those which could be removed without reducing

program effectiveness.

When the findings regarding the father-rated variables

and the homework/attendance data are considered jointly,

however, it becomes necessary to also consider another, more

conservative interpretation. These data suggest the

possibility that the specific components of the training

program, may not have been the primary factor in creating

group differences at posttreatment. One of the potential

difficulties in using a wait-list control design for

assessing treatment outcome is the difficulty in controlling

for non-specific factors such as receiving support,

expectations for change, or the influence of having invested

ones efforts in a treatment program (Basham, 1986). Looking

at the results from this standpoint, it could be argued that

improvement as rated by maternal reports was a function of

factors such as those listed. More particular to the

discrepancy between mother and father ratings of child

behavior problems, is the additional concern that parental










ratings of child behavior are social perceptions influenced

by parental characteristics as well as by objective

components of the child's behavior (e.g., Bates & Bayles,

1984). In fact, the influence of maternal characteristics,

especially as regards negative affect, on child behavior

ratings has been increasingly documented in the literature

on child psychopathology (e.g., Conrad & Hammen, 1989; Brody

& Forehand, 1986). It is feasible, therefore, that to the

extent that the program was successful in improving maternal

affect and psychosocial adjustment, changes in mother-rated

child behavior may be artifactual.

Taken together, these factors could explain the absence

of change in the reports of fathers who were not involved in

the program. It might also explain the absence of expected

relationships between level of participation and degree of

improvement, although it is possible that level of

participation could be expected to exert an influence on

non-specific factors listed above as well.

While a follow up study utilizing an active control

group, will be necessary to address this possibility more

definitively, several aspects of this study argue against

the influence of non-specific factors as an explanation for

mother-rated improvement. Firstly, maternal ratings did not

show a global pattern of improvement suggestive of a

positive response bias. As discussed earlier, the presence

of difficulties in the areas of relationship with spouse and

feeling restricted by the parenting role make sense in the










context of the research design and suggest that mothers were

likely accurately reporting their current feelings.

Moreover, given a bias toward positive reporting,

corresponding changes in ratings of child temperament would

have been expected; as discussed, these did not occur.

Finally, the PTTK scores indicated that improvement in the

active condition was accompanied by increases in knowledge

about child temperament and temperament-based approaches to

managing difficult child behavior. This suggests a link

between group participation and improvement.

This discussion of the observed results points to

several directions for continuing this research. An

immediate focus will be the examination of two month follow-

up data from this study. As described above in the section

regarding feelings of being restricted by the parenting

role, it is expected that the observed results may well be

stronger at follow up. The brief course of the training

program did not provide subjects with much time to practice

learned skills; it is expected that over time subjects will

have more opportunity to practice and may therefore

consolidate gains. This hypothesized improvement at follow-

up is, moreover, consistent with previous data regarding

outcome in child treatment which suggests that treatment

success looked at collectively may be stronger at follow-up

than at posttreatment (Levitt, 1957).

A related issue is that of the clinical impact of the

training program. As presented in the results section, the










attempt to provide information regarding the clinical

significance of the findings resulted in a somewhat mixed

picture. When examined at the group level, the results

suggest that the program was successful in bringing subjects

within the normal range; this was primarily an issue for the

Parenting Stress Index subscales as the majority of the

other scales were within normal limits prior to treatment.

However, at the individual level, the proportion of subjects

who moved into the normal range did not appear to be a

function of treatment. As was discussed above, it will be

necessary to examine the results at the follow-up to

determine whether a consolidation of gains is evident that

may impact the assessment of clinical significance. An

additional issue is that as the sample was not taken from a

clinic population, the definition of clinical significance

as reflecting movement into the normal range presented

methodological as well as conceptual difficulties. With

regard to the former, many subjects were within the normal

range at pretreatment; this limited the sample for examining

movement into the normal range. Many variables could not be

examined at a group level for the same reason. Similarly,

at the conceptual level it would seem that given a non-

clinic sample, an alternative conceptualization to clinical

significance may be beneficial. One possibility would be to

assess consumer satisfaction with the program (Kazdin,

1986). While this was not done in a standardized way in the










current study, it is a recommended approach for future

research.

Of primary importance in future projects will be the

inclusion of direct observations of parent-child

interactions. This will be necessary in order to assess

both parental use of program skills and potential changes in

child behavior. In particular, use of behavioral

observations would facilitate an examination of the extent

to which mother reports of improved child behavior tap

observable changes in the child as compared to reflecting

maternal bias due to improved affective state or investment

in the program as described earlier. The degree to which

mother-rated improvements in affect and comfort in the

parenting role are related to changes in child behavior and

maternal skill could also be examined. Finally, it would be

interesting to note whether more qualitative improvements in

the interaction such as maternal and child affect emerge at

posttreatment.

The inclusion of an alternative treatment condition

would seem potentially advantageous at the next stage in

this research program. Given the substantially larger

sample sizes needed to demonstrate differential

effectiveness between alternative treatments, and the

concomitantly larger investment in time and financial

resources, it has been suggested that individual approaches

be of demonstrated efficacy prior to engaging in comparative

outcome studies (Kazdin, 1986). The findings of this










initial study provide at least the preliminary support for

the parent training program necessary to warrant a more

intensive investigation. The advantage of a comparative

study would be to control for the effects of non-specific

factors discussed earlier. This would enable the

investigator to isolate the degree to which observed effects

are unique to the training program and potentially to more

closely link observed outcome to participation.

Future studies should also include fathers as

participating subjects. From a clinical perspective,

including fathers is congruent with feedback from group

participants as well as with the promotion of non-sexist and

collaborative parenting practices. From a research end,

including fathers would allow us to examine whether the

participation of both parents is related to increased

program efficacy. In particular, it would be interesting to

observe whether changes in spousal relationships occur as a

function of group participation.

Additionally, it may be useful to do an assessment at

mid-treatment after subjects have been familiarized with the

concept of temperament but before specific management

strategies have been addressed. Subjects in the current

study reported that it was an understanding of the

temperament characteristics rather than the specific

techniques which were most helpful. It would be useful to

determine the extent to which improvements can be attained

from that component of the program.









Finally, it would be advantageous to replicate this

study with a substantially larger sample. This would be

important in order to identify factors predictive of

treatment outcome. It is possible, for example, that the

parent-training program may be more effective given certain

child temperament characteristics or patterns relative to

others. Additionally, other factors such as subject

demographic characteristics or the specific form of child

behavior problems may influence outcome. Investigation of

these variables would facilitate continued program

development.















APPENDIX A
UNPUBLISHED MEASURES

Parent Test of Temperament Knowledge

Please circle the response which seems most correct to you.

1. Temperament refers to
a. An individual's style of behavior
b. A person's level of friendliness
c. An individual's degree of motivation
d. A person's tendency to act aggressively

2. It is most effective to
a. discuss with the child what the punishment should be
b. develop different discipline techniques to fit the
specific situation
c. have just a few discipline techniques that are used
consistently
d. just explain to the child what he or she has done
wrong

3. If one was to say that there was a "good fit" between a
parent and child, this would mean that
a. They love each other
b. There is a one-to-one agreement between the
temperament characteristics of the parent and those
of the child (for example, they are both active).
c. There is little or no stress and tension in the
relationship
d. The demands the parent places on the child are in
line with the child's abilities.

4. Children who demonstrate difficult temperament
characteristics
a. are usually hyperactive or emotionally disturbed
b. are generally perfectly normal, healthy children
c. are most frequently born prematurely
d. have generally suffered brain damage

5. The best way to prevent temper tantrums is to
a. Learn to recognize those situations which are
difficult for your child and develop ways to
decrease those demands
b. Avoid situations which are difficult for your child
c. Give in to your child's demands
d. Ignore them when they occur










6. The first step in learning to manage difficult child
behavior is to remain neutral. Which of the following is
involved in remaining neutral?
a. Focusing on the motivations underlying the child's
behavior
b. Focusing on your feelings in the situation
c. Focusing on the child's behavior
d. Reminding yourself that the child is acting
difficult in order to get your attention

7. When a child misbehaves it is important to
a. Tell him that he has broken the rule and describe
the punishment
b. Wait until later to punish him so that you have time
to decide what to do
c. Give him several warnings so that you don't punish
unfairly
d. Punish him quickly, so that explanations should be
given later.

8. Which of the following is true of temperament
characteristics?:
a. They are not influenced by the child's environment
b. They are primarily learned behaviors
c. They are relatively stable across time
d. They are not apparent until the child learns to
speak

9. Labeling the temperament characteristics that are related
to your child's behavior is one way to manage difficult
child behavior. When labeling for your child, it is
important
a. to use simple, specific language so that the child
can understand
b. to sound stern so that the child takes you seriously
c. to use many examples so that the child can
understand
d. to use an animated or emotional tone of voice so
that the child will pay attention to you

10. Which of the following is a way to prevent wild and
overactive behavior in an active, distractible child?
a. Have child play with only one toy or game so that he
or she doesn't start to run from one activity to
another
b. Provide child with a range of activities so that he
or she doesn't get bored with one toy.
c. Ignore the child when he or she starts to behave
wildly
d. Give the child extra attention when he or she starts
to behave wildly









11. The key to dealing with your child's unpredictable
(nonryhthmic) behavior is to
a. Avoid keeping the child on a set schedule
b. Try to keep a very routinized schedule until your
child gets used to it.
c. Let the child go to bed whenever he is tired
d. Separate bedtime from sleeptime and mealtime from
eating time

12. Which of the following is NOT an advantage of labeling
the temperament characteristics related to your child's
behavior?
a. It helps you see that the child is not being
intentionally hurtful.
b. It reminds you to remain neutral
c. It helps increase your child's understanding of his
or her behavior
d. It decreases the amount of difficult behavior

13. Which of the following has NOT been shown to be related
to having a child with difficult temperament
characteristics?
a. An increased feeling of attachment to the child
b. Feeling like one is a poor mother
c. Feeling restricted by the demands of parenting
d. Maternal depression

14. The goal of temperament-based parent training programs
is to
a. remove all stress and tension from the parent-child
relationship
b. improve the fit between the parents' demands and
child's characteristics.
c. teach parents to manage their children's behavior
without using punishment
d. help parents learn that it is sometimes necessary to
use stern punishments

15. Having a child teach a doll or stuffed animal how to
handle new situations is a good way to
a. Decrease distractibility
b. Help him or her overcome fears
c. Calm down an overactive child
d. Increase distractibility

16. When giving instructions to distractible children, it is
helpful to
a. Get directly to your point without wasting time
giving explanations or establishing eye contact
b. Give them a series of instructions while you have
their attention
c. Give one single-step command at a time
d. Sound stern so that the child knows your message is
important










17. Children who feel sad/angry frequently, may spend a
great deal of time complaining. One way to handle this
behavior is to
a. Ignore the child until he or she stops complaining
b. Let him continue to complain because it is important
that children know its O.K. to express their
feelings.
c. Tell the child that he should only speak when he has
something nice to say.
d. Acknowledge the child's feelings so that he knows
he's been heard but then firmly tell him that he has
a choice between talking nicely or staying quiet
until he feels better.

18. Some children have a hard time dealing with new
situations. With these children, it is helpful to
a. Briefly provide advance notice of the approaching
event so that they can get used to the idea
b. Avoid telling them about it so that they don't worry
in advance
c. Give them as much detail as possible about the
approaching event so that they know just what to
expect
d. Avoid demanding that they become involved in new
situations until they feel that they are ready

19. It is recommended that one uses "Cool Off" activities
a. To help a child deal with angry feelings
b. Before the child goes to bed.
c. When the child is becoming overexcited
d. To calm a child who is easily stimulated by sensory
information

20. If your child is prone to temper tantrums, it is
important to
a. Always ignore your child while he or she is
tantruming
b. Have a consistent way of responding to tantrums
c. Try a variety of approaches depending on the
situations
d. Avoid upsetting the child

21. Which of the following is NOT advisable when explaining
rules and consequences to children
a. Be serious in tone and manner when explaining rules
b. Explain rules to the child at a calm time.
c. Be simple and specific in explaining rules
d. Wait for the event to arise before discussing it.
Young children won't remember it otherwise









22. When parents attempt a new approach to managing
difficult child behavior, they should generally expect
a. Slow, steady, gradual improvements
b. Uneven progress marked by periods of improvement,
slips backward, and times without change
c. Slow, steady, gradual improvements except under
times of stress
d. Rapid improvement

23. If your child is experiencing difficulties in preschool
because of his or her temperament traits it is a good
idea to
a. Hope that his behavior in school will improve
because of the work you are doing at home
b. Take your child out of preschool until his or her
behavior is more manageable
c. Let the teacher know some of the techniques that
have worked for you at home
d. Switch schools until you find one that can better
manage his or her behavior.









Modified Impact on Family Scale

The questions on the following pages ask you to mark an
answer which best describes your feelings. While you may
not find an answer which exactly states your feelings,
please mark the answer which comes closest to describing how
you fel. YOUR FIRST REACTION TO EACH QUESTION SHOULD BE
YOUR ANSWER.

Please mark the degree to which you agree or disagree
with the following statements by choosing the number which
best matches how you feel:

Strongly Strongly
Disagree Disagree Not Sure Agree Agree

0 1 2 3 4

1. I spend a great deal of time with teachers and other
professionals concerning my child's behavior

2. I have had to cut down the hours I work in order to care
for my child.

3. I have stopped working because of my child's behavior.

4. Because of my child's behavior, our family is unable to
travel out of the city, or to other places we would like
to go.

5. It is hard to find a person who is willing or able to
take care of my child.

6. I see my relatives less than I would like because of my
child's behavior.

7. Because of my child's problems our family is not as close
as I would like.

8. I have thought about not having more children because of
the problems this child has had.

9. Because of my child's behavior I see friends less than I
would like.

10. It is hard to give much attention to my spouse/friend or
my other children because of the needs of my child.

11. Dealing with my child's problems has caused a great deal
of conflict between myself and my relatives.

12. Fatigue is a problem for me because of my child's
behavior.








82
13. Nobody understands the burden I carry because of my
child.

14. I worry about what will happen to my child in the
future (when he/she grows up).

15. There is much fighting within our family (eg., between
myself and my spouse, between our children) because of
my child's behavior.

16. Special family activities are often spoiled because of
my child's behavior.

17. There are few quiet, calm moments in our home because of
my child's behavior.

18. My child's behavior has caused conflicts between myself
and my neighbors.

19. It is hard for me to get my housework done when my
child is at home.

20. I am sometimes embarrassed by my child's behavior in
front of the neighbors.

21. It is often uncomfortable to have friends visit because
of my child's behavior.

22. My child's behavior places a financial burden on the
family.

23. For two parent homes:
My partner and I spend a great deal of our time together
talking about or dealing with my child's problems.

For single parent homes:
My child's behavior makes it difficult for me to enjoy
dating or socializing.
















APPENDIX B
RAW DATA


Pre-Treatment Assessment:
Parenting and Maternal Affect Va


riables


Treatment Condition

ID PCRELAT ATTACH


COMP

45
42
25
35
40
37
39
36
42
36
37
35
37
33
41
41
42
46
41
38
34


RESTRICT

31
24
16
27
26
14
26
28
26
18
20
31
23
21
26
29
24
28
30
23
17


DEPRESS

25
38
22
24
27
32
33
36
28
22
28
27
25
17
31
33
28
36
34
23
23


STAI

47
50
42
31
51
43
41
35
52
33
45
54
53
33
51
52
56
53
53
33
49


__________ _










Wait-list Condition


ATTACH COMP


PCRELAT

17
24
17
26
20
23
17
20
23
32
10
22
19
22
18
23
25
21
18


RESTRICT

23
27
22
23
24
31
20
24
34
33
26
26
20
21
19
27
15
34
20


DEPRESS

17
21
22
26
16
34
25
23
34
34
21
23
23
22
22
23
18
26
26


STAI

32
58
35
42
42
64
41
43
61
52
31
38
32
50
35
53
35

38










Pre-Treatment Assessment:
Child Behavior and Family Functioning Variables


Treatment Condition


INT EXT INT
ID MOM MOM DAD


EXT PDR
DAD MOM


PDR
DAD DISC


IFS SPREL


5 77 82
6 61 67 62
10 60 54 60
11 43 68 50 4
12 68 73
13 62 67 60 4
23 61 71 61 4
27 58 68 69
28 67 67 69 l
29 55 61 55
33 51 54 53
35 50 58
36 78 81 59
37 61 69 48
43 51 67 51
47 48 62
49 69 69 67
56 68 60 51
59 72 58 61
60 67 69 55
69 63 72 67

Wait-list Condition


INT EXT
ID MOM MOM

3 41 43
4 61 69
14 67 51
18 71 57
24 67 61
25 38 54
40 59 71
41 63 71
42 67 68
44 80 77
45 60 44
50 56 69
51 51 71
53 63 54
55 62 63
61 61 75
65 71 74
66 53 71
68 62 61


INT EXT
DAD DAD

48 46

42 33


33 34
64 81
63 64
61 68

53 39

33 55
72 57
66 71
60 73
71 73
45 64
51 50


18.60
15.60 20.60 12
8.30 9.60
10.30 14
8.60
9.60 7.60 16
12.30 14.30 16
11.67 4.00 17
8.67 12.33 16
16.33 11.00 13
15
10.33
17.00 14
12.33 13
6.00 12.00 11
12.30
16.33 14.00 13
8.33 6.00 14
13.33 9.33 12
10.66 6.33 14
15


PDR
MOM


PDR
DAD DISC IFS SPREL


3.33 11.60 19 27 16
6.60 7.30 57 24
3.30 4.30 12 27 24
11.60 30 19
16.30 35 26
11.00 4.00 12 37 17
11.00 12.00 15 41 21
14.33 21.50 17 35 16
18.67 11.67 17 67 26
15.00 77 28
6 39 15
13.00 46 27
15.00 2.00 16 61 23
15.33 14.67 14 27 26
12.50 9.00 13 36 19
13 43 27
S 16 39 14
17 52 25
9.67 6.67 12 40 21










Post-Treatment Assessment:
Parenting and Maternal Affect Variables


Treatment Condition


ID PCRELAT


ATTACH

12
14
14
15
11
15
10
12
16
13
10
11
14
11
13
16
16
17
17
17
12


COMP

31
31
32
33
41
36
33
33
44
34
30
29
35
25
33
32
27
41
34
32
26


RESTRICT

20
20
16
28
26
13
22
23
26
18
23
20
23
15
26
15
23
19
25
23
12


Wait-list Condition


ID PCRELAT


ATTACH COMP


RESTRICT

16
28
22
24
22
27
19
28
28
35
25
28
17
17
16
26
16
28
16


DEPRESS

15
24
20
18
23
20
26
26
20
20
24
16
22
12
24
27
22
18
22
18
14


STAI

33
40
30
37
48
44
41
36
43
31
44

44
22
46
42
32
42
50
30
31


DEPRESS

19
28
26
21
21
33
26
31
34
29
25
18
18
28
16
21
10
25
28


STAI

37
62
37
45
39
49
48
44
62
51
33
42
30

34
52
35
48
34










Post-Treatment Assessment:
Child Behavior and Family Functioning Variables


Treatment Condition


INT EXT INT EXT PDR
ID MOM MOM DAD DAD MOM


5 68
6 63
10 49
11 48
12 68
13 56
23 45
27 59
28 64
29 58
33 60
35 41
36 75
37 52
43 51
47 41
49 53
56 63
59 57
60 45
69 41


PDR
DAD DISC IFS SPREL


63 6.00 37 22
68 67 74 11.67 13.67 11 43 24
48 60 55 4.67 8.00 22
68 58 67 14.67 10 51 16
71 7.67 34
57 51 51 5.67 6.00 15 27 15
48 38 48 7.67 9.50 15 32 24
68 72 72 5.67 4.00 15 53 16
64 71 71 6.67 14.33 14 45 26
50 10.33 6.00 13 26
59 56 48 15 20 17
48 45
79 52 63 14.33 15.00 10 61 29
65 58 61 10.00 10 11 16
65 42 50 8.33 4.00 12 27 24
51 8.67 6 19
54 59 51 8.30 6.50 11 36 26
50 59 54 5.67 3.67 13 19 24
47 51 50 8.00 3.33 10 28 26
54 52 54 5.00 8.00 13 15 22
46 59 67 14 26 19


Wait-list Condition


INT EXT
ID MOM MOM

3 .
4 61 68
14 66 52
18 66 52
24 60 55
25 51 51
40 53 68
41 61 71
42 58 67
44 83 75
45 70 47
50 51 67
51 48 60
53 60 51
55 64 71
61 61 74
65 70 67
66 58 75
68 62 64


INT EXT PDR
DAD DAD MOM


PDR
DAD DISC IFS SPREL


1.33 3.67 30 13
13.33 12.00 69 26
39 36 5.00 1.67 12 33 30
8.67 25 17
19.00 27 24
S 6.67 2.67 36 18
67 77 12.33 11.50 17 30 22
58 68 17.00 16.00 19 46 23
56 64 9.33 15 48 34
69 68 15 74 29
51 39 6 30 15
S 8.00 49 30
12.33 5.50 45 18
68 54 16.67 12.00 13 17 25
61 64 12.00 8.00 12 28 24
60 68 12 31 21
40 52 12 14 11
48 62 15 64 25
45 39 8.00 10.67 11 39 24










Pre-Treatment Assessment: Temperament Variables


Condition


ID ACTV ADAPT APPRO DISTR INTEN MOOD


4.63
3.50
2.88
5.63
4.50
3.13
4.13
3.88
3.75
3.63
3.38
4.38
3.50
4.38
4.68
4.00
4.25
3.13
2.38
4.25
4.50


4.25
2.50
2.50
3.75
1.50
3.00
2.50
4.38
2.50
3.00
2.88
2.29
1.63
3.63
4.38
4.00
2.88
2.38
3.38
2.75
2.13


4.88
3.75
6.38
2.50
6.25
1.38
1.50
3.63
1.13
2.00
2.63
0.75
2.13
2.88
4.25
5.13
0.88
0.88
1.88
2.00
3.88


2.50
2.38
3.25
2.00
1.88
2.88
1.63
3.63
1.75
1.88
2.88
2.25
1.75
2.25
2.25
0.88
1.63
2.63
2.38
2.00
2.63


4.75
5.13
4.38
4.63
5.63
3.63
5.50
4.38
3.63
4.38
4.50
4.75
5.50
4.63
4.75
5.50
5.13
4.63
4.63
5.25
5.13


3.50
3.50
3.63
4.00
3.00
3.13
3.13
2.88
2.63
3.57
3.00
1.38
2.25
3.13
3.75
4.25
3.25
2.88
3.13
2.38
3.50


PERS RHYTHM THRESH


5.25
3.75
3.63
3.75
3.50
3.75
4.00
3.63
3.13
2.88
3.38
3.50
3.75
2.88
3.00
4.88
3.50
3.88
3.00
1.25
2.25


3.13
2.00
4.63
4.63
5.00
5.00
5.25
3.13
3.50
2.63
3.00
2.13
4.25
3.75
3.88
4.13
4.75
4.00
2.28
4.25
2.75


4.75
2.25
2.50
3.63
2.50
3.25
1.88
2.38
2.75
3.25
1.38
2.38
2.13
1.38
3.63
2.25
2.63
3.63
1.00
3.75
2.50


Wait-list


Condition


ID ACTV ADAPT APPRO DISTR INTEN MOOD


3.63
5.50
1.75
3.00
3.88
3.63
4.88
4.13
3.50
5.63
3.88
3.50
4.25
3.63
3.50
4.38
3.88
3.63
3.50


3.72
1.88
3.25
3.63
2.75
2.50
4.00
2.88
1.25
1.86
1.38
3.50
4.50
3.25
3.13
3.50
3.00
3.13
3.13


4.00
2.38
2.38
3.63
2.25
2.13
3.13
1.88
0.88
2.25
1.50
2.75
3.88
4.00
4.63
3.25
3.63
1.88
1.75


2.88
3.38
3.50
2.13
2.50
2.75
2.38
1.50
2.13
0.38
3.75
2.88
3.13
3.50
2.88
2.00
2.00
2.38
2.63


3.00
5.50
4.13
3.88
5.00
4.38
3.63
4.75
6.00
6.00
3.38
3.75
4.50
4.63
3.63
5.13
5.25
4.81
4.63


4.28
2.88
3.38
2.71
4.13
3.25
5.13
2.25
3.57
1.88
3.25
4.50
3.38
4.38
2.57
4.13
2.88
4.00
3.50


PERS RHYTHM THRESH

2.25 2.07 1.500
2.88 2.63 3.885
3.25 3.75 1.885
4.00 4.14 0.885
3.00 3.13 0.385
3.13 3.75 2.500
3.13 2.75 4.000
3.63 4.38 1.500
3.75 3.38 0.750
3.38 1.25 2.250
2.88 4.13 0.635
2.88 2.63 2.385
3.63 4.00 2.750
3.75 3.63 2.250
2.88 3.88 2.630
3.88 4.38 3.630
3.75 3.43 1.500
4.00 3.75 3.000
3.63 4.75 0.880


_LLS~U~rfI%"A i t- i nn


--- ---- ----~----~-


Treatment










Post-Treatment Assessment: Temperament Variables

Treatment Condition

ID ACTV ADAPT APPRO DISTR INTEN MOOD PERS RHYTHM THRESH

5 4.62 4.25 4.87 2.50 4.75 3.50 5.25 3.12 4.75
6 3.50 2.50 3.75 2.37 5.12 3.50 3.75 2.00 2.25
10 2.87 2.50 6.37 3.25 4.37 3.62 3.62 4.62 2.50
11 5.62 3.75 2.50 2.00 4.62 4.00 3.75 4.62 3.62
12 4.50 1.50 6.25 1.87 5.62 3.00 3.50 5.00 2.50
13 3.12 3.00 1.37 2.87 3.62 3.12 3.75 5.00 3.25
23 4.12 2.50 1.50 1.62 5.50 3.12 4.00 5.25 1.87
27 3.87 4.37 3.62 3.62 4.37 2.87 3.62 3.12 2.37
28 3.75 2.50 1.13 1.75 3.63 2.63 3.13 3.50 2.75
29 3.62 3.00 2.00 1.87 4.37 3.57 2.87 2.62 3.25
33 3.37 2.87 2.62 2.87 4.50 3.00 3.37 3.00 1.37
35 4.37 2.28 0.75 2.25 4.75 1.37 3.50 2.12 2.37
36 3.50 1.62 2.12 1.75 5.50 2.25 3.75 4.25 2.12
37 4.37 3.62 2.87 2.25 4.62 3.12 2.87 3.75 1.37
43 4.62 4.37 4.25 2.25 4.75 3.75 3.00 3.87 3.62
47 4.00 4.00 5.12 0.87 5.50 4.25 4.87 4.12 2.25
49 4.25 2.87 0.87 1.62 5.12 3.25 3.50 4.75 2.62
56 3.13 2.38 0.88 2.63 4.63 2.88 3.88 4.00 3.63
59 2.38 3.38 1.88 2.38 4.63 3.13 3.00 2.28 1.00
60 4.25 2.75 2.00 2.00 5.25 2.38 1.25 4.25 3.75
69 4.50 2.13 3.88 2.63 5.13 3.50 2.25 2.75 2.50

Wait-list Condition

ID ACTV ADAPT APPRO DISTR INTEN MOOD PERS RHYTHM THRESH

3 3.62 3.71 4.00 2.87 3.00 4.28 2.25 2.07 1.50
4 5.50 1.87 2.37 3.37 5.50 2.87 2.87 2.62 3.87
14 1.75 3.25 2.37 3.50 4.12 3.37 3.25 3.75 1.87
18 3.00 3.62 3.62 2.12 3.87 2.71 4.00 4.14 0.87
24 3.87 2.75 2.25 2.50 5.00 4.12 3.00 3.12 0.37
25 3.62 2.50 2.12 2.75 4.37 3.25 3.12 3.75 2.50
40 4.87 4.00 3.12 2.37 3.62 5.12 3.12 2.75 4.00
41 4.12 2.87 1.87 1.50 4.75 2.25 3.62 4.37 1.50
42 3.50 1.25 0.87 2.12 6.00 3.57 3.75 3.37 0.75
44 5.62 1.85 2.25 0.37 6.00 1.87 3.37 1.25 2.25
45 3.87 1.37 1.50 3.75 3.37 3.25 2.87 4.12 0.62
50 3.50 3.50 2.75 2.87 3.75 4.50 2.87 2.62 2.37
51 4.25 4.50 3.87 3.12 4.50 3.37 3.62 4.00 2.75
53 3.63 3.25 4.00 3.50 4.63 4.38 3.75 3.63 2.25
55 3.50 3.13 4.63 2.88 3.63 2.57 2.88 3.88 2.63
61 4.38 3.50 3.25 2.00 5.13 4.13 3.88 4.38 3.63
65 3.88 3.00 3.63 2.00 5.25 2.88 3.75 3.43 1.50
66 3.63 3.13 1.88 2.38 4.81 4.00 4.00 3.75 3.00
68 3.50 3.13 1.75 2.63 4.63 3.50 3.63 4.75 0.88










Post-Treatment Assessment: Temperament Variables


Treatment Condition


ID ACTV ADAPT APPRO DISTR INTEN MOOD


4.25 4.87
2.50 3.75
2.50 6.37
3.75 2.50
1.50 6.25
3.00 1.37
2.50 1.50
4.37 3.62
2.50 1.13
3.00 2.00
2.87 2.62
2.28 0.75
1.62 2.12
3.62 2.87
4.37 4.25
4.00 5.12
2.87 0.87
2.38 0.88
3.38 1.88
2.75 2.00
2.13 3.88


2.50
2.37
3.25
2.00
1.87
2.87
1.62
3.62
1.75
1.87
2.87
2.25
1.75
2.25
2.25
0.87
1.62
2.63
2.38
2.00
2.63


4.75
5.12
4.37
4.62
5.62
3.62
5.50
4.37
3.63
4.37
4.50
4.75
5.50
4.62
4.75
5.50
5.12
4.63
4.63
5.25
5.13


3.50
3.50
3.62
4.00
3.00
3.12
3.12
2.87
2.63
3.57
3.00
1.37
2.25
3.12
3.75
4.25
3.25
2.88
3.13
2.38
3.50


PERS RHYTHM THRESH


5.25 3.12
3.75 2.00
3.62 4.62
3.75 4.62
3.50 5.00
3.75 5.00
4.00 5.25
3.62 3.12
3.13 3.50
2.87 2.62
3.37 3.00
3.50 2.12
3.75 4.25
2.87 3.75
3.00 3.87
4.87 4.12
3.50 4.75
3.88 4.00
3.00 2.28
1.25 4.25
2.25 2.75


Wait-list Condition


ID ACTV

3 3.62
4 5.50
14 1.75
18 3.00
24 3.87
25 3.62
40 4.87
41 4.12
42 3.50
44 5.62
45 3.87
50 3.50
51 4.25
53 3.63
55 3.50
61 4.38
65 3.88
66 3.63
68 3.50


ADAPT APPRO DISTR INTEN MOOD


3.71
1.87
3.25
3.62
2.75
2.50
4.00
2.87
1.25
1.85
1.37
3.50
4.50
3.25
3.13
3.50
3.00
3.13
3.13


4.00
2.37
2.37
3.62
2.25
2.12
3.12
1.87
0.87
2.25
1.50
2.75
3.87
4.00
4.63
3.25
3.63
1.88
1.75


2.87
3.37
3.50
2.12
2.50
2.75
2.37
1.50
2.12
0.37
3.75
2.87
3.12
3.50
2.88
2.00
2.00
2.38
2.63


3.00 4.28
5.50 2.87
4.12 3.37
3.87 2.71
5.00 4.12
4.37 3.25
3.62 5.12
4.75 2.25
6.00 3.57
6.00 1.87
3.37 3.25
3.75 4.50
4.50 3.37
4.63 4.38
3.63 2.57
5.13 4.13
5.25 2.88
4.81 4.00
4.63 3.50


PERS RHYTHM THRESH


2.25
2.87
3.25
4.00
3.00
3.12
3.12
3.62
3.75
3.37
2.87
2.87
3.62
3.75
2.88
3.88
3.75
4.00
3.63


2.07
2.62
3.75
4.14
3.12
3.75
2.75
4.37
3.37
1.25
4.12
2.62
4.00
3.63
3.88
4.38
3.43
3.75
4.75


1.50
3.87
1.87
0.87
0.37
2.50
4.00
1.50
0.75
2.25
0.62
2.37
2.75
2.25
2.63
3.63
1.50
3.00
0.88


4.62
3.50
2.87
5.62
4.50
3.12
4.12
3.87
3.75
3.62
3.37
4.37
3.50
4.37
4.62
4.00
4.25
3.13
2.38
4.25
4.50


4.75
2.25
2.50
3.62
2.50
3.25
1.87
2.37
2.75
3.25
1.37
2.37
2.12
1.37
3.62
2.25
2.62
3.63
1.00
3.75
2.50











PCRELAT



ATTACH

COMP

RESTRICT

DEPRESS

STAI

INT MOM



EXT MOM



PDR MOM

INT DAD



EXT DAD



PDR DAD

DISC



IFS

SPREL

ACTV

ADAPT

APPRO

DISTR

INTEN


91
Index of Variable Names and Codes

Parent-Child Relationship (Cleminshaw

Guidubaldi Parent Satisfaction Scale)

Parent Attachment (PSI)

Parent Sense of Competence (PSI)

Restrictions Imposed by Parental Role (PSI)

Parent Depression (PSI)

Trait Anxiety (State Trait Anxiety Inventory)

Mother-rated Internalizing Behavior Problems

(CBCL)

Mother-rated Externalizing Behavior Problems

(CBCL)

Mother-rated Parent Daily Report

Father-rated Internalizing Behavior Problems

(CBCL)

Father-rated Externalizing Behavior Problems

(CBCL)

Father-rated Parent Daily Report

Family Discipline and Control (Cleminshaw

Guidubaldi Parent Satisfaction Scale)

Modified Impact on Family Scale

Relationship with Spouse (PSI)

Activity Level

Adaptability

Approach vs Withdrawal

Distractibility

Intensity










MOOD Mood

PERS Persistance

RHYTHM Rhythmicity/Regularity

THRESH Threshold of Responsiveness

AGE Child Age

MAGE Maternal Age

PAGE Paternal Age

MED Maternal Education (in years)

PED Paternal Education (in years)

SIB Number of siblings

INC Income (in thousands of dollars)

MS Maternal Marital Status (l=Married;

2=Divorced/Separated; 3=single; 4=Living with

child's father)

RACE Maternal Race (l=White; 2=Black; 3=Mixed)

SEX Child's Sex (l=Male; 2=Female)















APPENDIX C
Temperament-Focused Parent-Training:
Outline of a Psychoeducational Program for
Parents with Temperamentally Difficult Preschoolers

Overview of Parent-Training Program

The following outline provides a detailed description
of a training program for parents of temperamentally
difficult preschool children. Based on the goodness-of-fit
notion, the training is geared toward utilizing an
understanding of child temperament to enhance the match
between child characteristics and parental demands. The
program is largely based on the approach developed by
Stanley Turecki, a child psychiatrist at Beth Israel
Hospital in New York, and presented in The Difficult Child
(Turecki & Tonner, 1985). However, his program has been
modified both to include recommendations offered in the
works of Thomas and Chess and to reflect my understanding of
child temperament and behavior management.
The initial phase of the program is geared towards
familiarizing parents with the nature of child temperament
and its role in understanding child behavior. This is
addressed as regards understanding both children in general,
and the difficulties experienced in one's own family. An
emphasis will be placed on understanding the specific ways
that temperament can be used to describe an individual
child's behavior. For instance, an "unadaptable" child may
be hesitant with new people, but love new and different
toys.
The second phase involves the acquisition of specific
management techniques for a range of behaviors associated
with temperament characteristics as well as the development
of a consistent approach to problematic, non-temperament
behaviors. All of the techniques presented are
conceptualized as general suggestions which are not
expected to suit everyone. Additionally, these suggestions
in no way exhaust the available options. Group members are
advised to make changes in accordance with their own needs
and styles so long as they stay within the conceptual
framework of the goodness-of-fit model. Members are
encouraged to discuss proposed modifications both to
facilitate sharing of ideas among group participants and to
check that approaches used by group members are congruent
with this model.









Session #1 Introduction to Temperament

A. Opening

1. Introduce self and give brief overview of program

2. Have group members introduce themselves and speak
for a short time about their reasons for coming for
the program

B. Temperament

1. Introduce idea of temperament by reading a couple of
vignettes describing children with combinations of
temperamental traits which can at times make them
difficult to parent.
a. Ask group about any of these characteristics
which they see in their children. As they
discuss the behaviors they see, write the
relevant temperament traits on the board.

2. Temperament: Refers to the "how" or the style of
behavior. It is different from ability, which is
the "how well" of behavior or motivation, which is
the "why".
a. Ex. Two children may choose to play the same
game, and play it equally well and still differ
in their styles. For instance: in how much
they run around; how upset they are if the
rules are changed; or how easily they can leave
the game if you call them in for dinner.
b. Temperament traits are fairly consistent ways
of responding to situations and appear to be
largely inborn in that these styles are
apparent very early.

3. Brief History: Work in this area was begun
approximately 30 years ago by two child
psychiatrists, Thomas and Chess. They began their
work because they were dissatisfied with the
predominant view that children are born as "blank
slates", such that everything they become is the
responsibility of the parents, teachers, and child
care professionals.
a. Saw differences in children too early to
attribute them to variability in parental
handling
b. As clinicians they were aware that not all
behavior problems could be related to parental
behaviors
c. Objected to the tendency which grew out of a
"blank slate" approach to blame the mother for
all behavioral or psychological problems









4. Identified Nine Dimensions of Temperament: As go
over these refer to list of behaviors that parents
identified earlier. Give handout on temperament
dimensions
a. Activity Level
b. Regularity in biological functioning:
Predictability in patterns of sleep and hunger
c. Approach vs Withdrawal to new situations
d. Adaptability: Ease with which child comes to
adjust to new or changed situation after the
initial response
e. Threshold of Response to sensory stimulation:
sensitivity to sensory stimulation such as the
feel of materials, bright lights, texture of
food, noise
f. Mood: amount of pleasant, happy, friendly
behavior vs crying, complaining, unpleasant
behavior
g. Intensity of reaction: energy level of response
h. Distractibility: How easily is the child
distracted from what he/she is doing.
i. Persistence: Amount of time the child will stay
with one activity, particularly in the face of
events that make it difficult to continue.

5. When children display behaviors that are at one
extreme of many of these dimensions, they can be
difficult to parent as was the case in the vignettes
we read earlier. Each of you were invited to join
the group both because your child demonstrates some
of these behaviors and because you reported finding
that these behavioral styles make parenting more
difficult for you.
a. However, each of these temperament
characteristics that make the child seem
difficult in one situation can be an asset in a
different situation or stage of development.
b. Ask group to identify how each temperament
trait may be easy or difficult depending on the
situation. Supply examples that they miss.

6. These children, even at their most difficult, are
normal. Demonstrating temperamental traits that can
make parenting difficult does mean that the child is
emotionally disturbed, brain damaged, hyperactive,
or in anyway "abnormal". Relatedly, it also doesn't
mean that you've been a bad parent.

7. Children with traits that can make parenting more
difficult are not doing so on purpose. Their
temperament make-up means that a variety of
situations are harder or less enjoyable for them
than they may be for children with a different set
of characteristics. As their parents, you bear the